Taking Your Talent To The Web

00 0732 FM 4/24/01 1:38 PM Page i
Taking Your
Talent to the
Web
A Guide for the Transitioning Designer
By Jeffrey Zeldman
201 West 103rd Street, Indianapolis, Indiana 46290
PDF compression, OCR, web optimization using a watermarked evaluation copy of CVISION PDFCompressor
00 0732 FM 4/24/01 1:38 PM Page ii
Taking Your Talent to the Web:

Publisher
A Guide for the Transitioning Designer David Dwyer
Copyright  2001 by New Riders Publishing Associate Publisher
Al Valvano
All rights reserved. No part of this book shall be reproduced, stored in
a retrieval system, or transmitted by any means—electronic, mechani- Executive Editor
cal, photocopying, recording, or otherwise—without written permission Karen Whitehouse
from the publisher. No patent liability is assumed with respect to the Acquisitions Editor
use of the information contained herein. Although every precaution Michael Nolan
has been taken in the preparation of this book, the publisher and au-
thor assume no responsibility for errors or omissions. Neither is any li- Technical Editor
Steve Champeon
ability assumed for damages resulting from the use of the information
contained herein. Development Editor
Victoria Elzey
International Standard Book Number: 0-7357-1073-2
Product Marketing
Library of Congress Catalog Card Number: 00-111152 Manager
Kathy Malmloff
Printed in the United States of America
Managing Editor
First Printing: May 2001 Sarah Kearns
Project Editor
05 04 03 02 01 7 6 5 4 3 2 1
Jake McFarland
Interpretation of the printing code: The rightmost double-digit number Copy Editor
is the year of the book’s printing; the rightmost single-digit number is Chrissy Andry
the number of the book’s printing. For example, the printing code 01-
Cover Designer
1 shows that the first printing of the book occurred in 2001.
Allison Cecil
Trademarks Interior Designer

All terms mentioned in this book that are known to be trademarks or Suzanne Pettypiece
service marks have been appropriately capitalized. New Riders Publish- Compositor
ing cannot attest to the accuracy of this information. Use of a term in Suzanne Pettypiece
this book should not be regarded as affecting the validity of any trade-
mark or service mark. Proofreader
Jeannie Smith
Warning and Disclaimer Indexers
Every effort has been made to make this book as complete and as ac- Lisa Stumpf
curate as possible, but no warranty of fitness is implied. The informa- Larry Sweazy
tion provided is on an “as is” basis. The authors and the publisher shall
have neither liability nor responsibility to any person or entity with re-
spect to any loss or damages arising from the information contained in
this book.
00 0732 FM 4/24/01 1:57 PM Page iii
Contents at a Glance
Introduction 1
Part I WHY: Understanding the Web
Chapter 1 Splash Screen 5
Chapter 2 Designing for the Medium 13
Chapter 3 Where Am I? Navigation & Interface 69
Part II WHO: People, Parts, and Processes
Chapter 4 How This Web Thing Got Started 111
Chapter 5 The Obligatory Glossary 123
Chapter 6 What Is a Web Designer, Anyway? 135
Chapter 7 Riding the Project Life Cycle 147
Part III HOW: Talent Applied (Tools & Techniques)
Chapter 8 HTML, the Building Blocks of Life Itself 175
Chapter 9 Visual Tools 209
Chapter 10 Style Sheets for Designers 253
Chapter 11 The Joy of JavaScript 285
Chapter 12 Beyond Text/Pictures 327
Chapter 13 Never Can Say Goodbye 387
Index 403
00 0732 FM 4/24/01 1:57 PM Page iv
Table of Contents
Introduction 1
Part I WHY: Understanding the Web 3
1 Splash Screen 5
Meet the Medium 6
Expanding Horizons 7
Working the Net…Without a Net 9
Smash Your Altars 11
2 Designing for the Medium 13

Breath Mint? Or Candy Mint? 14
Where’s the Map? 19
Mars and Venus 20
Web Physics: Action and Interaction 20
Different Purposes, Different Methodologies 23
Web Agnosticism 23
Open Standards—They’re Not Just for Geeks Anymore 27
Point #1: The Web Is Platform-Agnostic 27
Point #2: The Web Is Device-Independent 29
Point #3: The Web Is Held Together by Standards 29
The 18-Month Pregnancy 31
Chocolatey Web Goodness 32
’Tis a Gift to Be Simple 32
Democracy, What a Concept 32
Instant Karma 34
The Whole World in Your Hands 35
Just Do It: The Web as Human Activity 35
The Viewer Rules 36
Multimedia: All Talking! All Dancing! 37
The Server Knows 38
It’s the Bandwidth, Stupid 41
Web Pages Have No Secrets 42
The Web Is for Everyone! 44
iv
00 0732 FM 4/24/01 1:38 PM Page v
It’s Still the Bandwidth, Stupid 45

Swap text and code for images 46
Trim those image files 46
Do more with less 47
Prune redundancy 47
Cache as Cache Can 49
Much Ado About 5K 50
Screening Room 51
Liquid Design 51
Color My Web 55
Thousands Weep 57
Gamma Gamma Hey! 59
Typography 62
The 97% Solution 62
Points of Distinction 63
Year 2000—Browsers to the Rescue 64
Touch Factor 65
Appropriate Graphic Design 65
Accessibility, the Hidden Shame
of the Web 65
User Knowledge 67
3 Where Am I? Navigation & Interface 69

What Color Is Your Concept? 70
Business as (Cruel and) Usual 71
The Rise of the Interface Department 72
Form and Function 74
Copycats and Pseudo-Scientists 77
Chaos and Clarity 78
A Design Koan: Interfaces Are a Means too Often Mistaken for an End 80
Universal Body Copy and Other Fictions 80
Interface as Architecture 81
Ten (Okay, Three) Points of Light 82
Be Easily Learned 82
Remain Consistent 82
Continually Provide Feedback 84
GUI, GUI, Chewy, Chewy 84
It’s the Browser, Stupid 85
00 0732 FM 4/24/01 1:38 PM Page vi
Clarity Begins at Home (Page) 87

I Think Icon, I Think Icon 88
Structural Labels: Folding the Director’s Chair 90
The Soul of Brevity 90
Hypertext or Hapless Text 91
Scrolling and Clicking Along 95
Stock Options (Providing Alternatives) 97
Hierarchy and the So-Called Three Click Rule 97
The So-Called Rule of Five 99
Highlights and Breadcrumbs 101
Consistent Placement 102
Brand That Sucker! 103
Part II WHO: People, Parts, and Processes 109
4 How This Web Thing Got Started 111

1452 111
1836 111
1858 111
1876 112
Why We Mentioned These Things 112
1945 112
1962 112
1965 112
1966 113
1978 113
1981 113
1984 114
1986 114
1988 114
1989 115
1990 115
1991 115
1993 116
1994 116
1995 117
1996 118
1997 119
1998 120
1999 121
vi
00 0732 FM 4/24/01 1:38 PM Page vii
2000 121
The year web standards broke, 1 121
The year the bubble burst 122
2001 122
5 The Obligatory Glossary 123

Web Lingo 124
Extranet 124
HTML 125
Hypertext, hyperlinks, and links 125
Internet 125
Intranet 126
JavaScript, ECMAScript, CSS, XML, XHTML, DOM 127
Web page 128
Website 128
Additional terminology 129
Roles and Responsibilities in the Web World 129
Web developer/programmer 129
Project manager 130
Systems administrator (sysadmin) and network administrator (netadmin) 131
Web technician 131
Your Role in the Web 133
6 What Is a Web Designer, Anyway? 135

What We Have Here Is an Opportunity to Communicate 137
The Definition Defined 138
Look and feel 138
Business-to-business 139
Business-to-consumer 140
Solve Communication Problems 140
Brand identity 141
Web-specific 141
Restrictions of the Medium 142
Technology 143
Works with team members 144
Visually and emotionally engaging 144
Easy to navigate 145
Compatible with visitors’ needs 145
Accessible to a wide variety of web browsers and other devices 145
Can You Handle It? 146
vii
00 0732 FM 4/24/01 1:38 PM Page viii
7 Riding the Project Life Cycle 147

What Is the Life Cycle? 148
Why Have a Method? 149
We Never Forget a Phase 151
Analysis (or “Talking to the Client”) 152
The early phase 153
Defining requirements 154
Design 156
Brainstorm and problem solve 156
Translate needs into solutions 157
Sell ideas to the client 158
Identify color comps 160
Create color comps/proof of concept 160
Present color comps and proof of concept 161
Receive design approval 162
Development 162
Create all color comps 163
Communicate functionality 164
Work with templates 165
Design for easy maintenance 165
Testing 166
Deployment 166
The updating game 167
Create and provide documentation and style guides 168
Provide client training 169
Learn about your client’s methods 169
Work the Process 170
Part III HOW: Talent Applied (Tools & Techniques) 173

8 HTML, the Building Blocks of Life Itself 175
Code Wars 176
Table Talk 176
XHTML Marks the Spot 177
Minding Your <p>’s and q’s 178
Looking Ahead 179
Getting Started 181
View Source 183
A Netscape Bonus 184
The Mother of All View Source Tricks 184
viii
00 0732 FM 4/24/01 1:38 PM Page ix
Doin’ it in Netscape 184

Doin’ it in Internet Explorer 185
Absolutely Speaking, It’s All Relative 185
What Is Good Markup? 188
What Is Sensible Markup? 189
HTML as a Design Tool 190
Plug-ins and Tables and Frames, Oh My! 193
The Frames of Hazard 194
Please Frame Safely 195
Framing Your Art 195
<META> <META> Hiney Ho! 197
Search Me 197
Take a (Re)Load Off 200
A Comment About <COMMENTS> 201
WYSIWYG, My Aunt Moira’s Left Foot 202
Code of Dishonor 202
WYS Is Not Necessarily WYG 203
Browser Incompatibilities: Can’t We All Just Get Along? 204
Publish That Sucker! 205
HTMHell 207
9 Visual Tools 209

Photoshop Basics: An Overview 209
Comp Preparation 210
Dealing with Color Palettes 210
Exporting to Web-Friendly Formats 210
Gamma Compensation 211
Preparing Typography 211
Slicing and Dicing 211
Rollovers (Image Swapping) 211
GIF Animation 212
Create Seamless Background Patterns (Tiles) 212
Color My Web: Romancing the Cube 212
Dither Me This 213
Death of the Web-Safe Color Palette? 215
A Hex on Both Your Houses 216
Was Blind, but Now I See 217
From Theory to Practice 217
Format This: GIFs, JPEGs, and Such 221
ix
00 0732 FM 4/24/01 1:38 PM Page x
GIF 222
Loves logos, typography, and long walks in the woods 223
GIFs in Photoshop 224
JPEG, the Other White Meat 226
Optimizing GIFs and JPEGs 228
Expanding on Compression 231
Make your JPEGS smaller 231
Combining sharp and blurry 231
Compression Breeds Style: Thinking About the Medium 234
PNG 236
Animated GIFs 237
Creating Animations in ImageReady 238
Typography 239
The ABCs of Web Type 240
Anti-Aliasing 241
Specifying Anti-Aliasing for Type 243
General tips 244
General Hints on Type 244
The Sans of Time 244
Space Patrol 245
Lest We Fail to Repeat Ourselves 245
Accessibility, Thy Name Is Text 246
Navigation: Charting the Visitor’s Course 247
Slicing and Dicing 248
Thinking Semantically 251
10 Style Sheets for Designers 253

Tag Soup and Crackers 254
CSS to the Rescue…Sort of 256
Designing with Style: Cascading Style Sheets (CSS) 257
Separation of Style from Content 258
Disadvantages of Traditional Web Design Methods 258
CSS Advantages: Short Term 259
CSS Advantages: Long Term 261
Compatibility Problems: An Overview 261
Working with Style Sheets 263
Types of Style Sheets 266
External style sheets 267
Embedding a style sheet 268
Adding styles inline 269
00 0732 FM 4/24/01 1:38 PM Page xi
Trouble in Paradise: CSS Compatibility Issues 271

Fear of Style Sheets: CSS and Layout 271
Fear of Style Sheets: Leading and Image Overlap 273
Fear of Style Sheets: CSS and Typography 274
Promise and performance 274
Font Size Challenges 276
Points of contention 276
Point of no return: browsers of the year 2000 277
Pixels for fun and profit 278
Absolute size keywords 280
Relative keywords 281
Length units 282
Percentage units 283
Looking Forward 284
11 The Joy of JavaScript 285

What Is This Thing Called JavaScript? 286
The Web Before JavaScript 286
JavaScript, Yesterday and Today 287
JavaScript, Unhh! What Is It Good For? 288
Sounds Great, but I’m an Artist. Do I Really Have to Learn This Stuff? 290
Educating Rita About JavaScript 291
Don’t Panic! 292
JavaScript Basics for Web Designers 292
The Dreaded Text Rollover 294
The Event Handler Horizon 295
Status Quo 297
A Cautionary Note 299
Kids, Try This at Home 299
The fine print 299
Return of the son of fine print 300
The Not-So-Fine Print 300
The Ever-Popular Image Rollover 302
A Rollover Script from Project Cool 303
Windows on the World 307
Get Your <HEAD> Together 308
Avoiding the Heartbreak of Linkitis 310
Browser Compensation 312
JavaScript to the Rescue! 314
Location, location, location 315
xi
00 0732 FM 4/24/01 1:38 PM Page xii
Watching the Detection 316

Going Global with JavaScript 321
Learning More 324
12 Beyond Text/Pictures 327

Prelude to the Afternoon of Dynamic Websites 329
You Can Never Be Too Rich Media 330
The Form of Function: Dynamic Technologies 330
Server-Side Stuff 331
Where were you in ‘82? 332
Indiana Jones and the template of doom 332
Serving the project 334
Doing More 335
Mini-Case Study: Waferbaby.com 336
Mini-Case Study: Metafilter.com 337
Any Size Kid Can Play 338
Take a Walk on the Server Side 339
Are You Being Served? 341
Advantages of SSI 342
Disadvantages of SSI 343
Cookin’ with Java 343
Ghost in the Virtual Machine 344
Where the web designer fits in 346
Java Woes 347
Java Woes: The Politically Correct Version 347
Java Joys 349
Rich Media: Exploding the “Page” 350
Virtual Reality Modeling Language (VRML) 350
SVG and SMIL 352
SMIL (through your fear and sorrow) 352
SVG for You and Me 354
Romancing the logo 356
Sounds dandy, but will it work? 357
Promises, Promises 358
Turn on, Tune in, Plug-in 358
A Hideous Breach of Reality 360
The ubiquity of plug-ins 360
xii
00 0732 FM 4/24/01 1:38 PM Page xiii
The Impossible Lightness of Plug-ins 361

Plug-ins Most Likely to Succeed 361
RealPlayer (www.real.com) 362
QuickTime (www.apple.com/quicktime/) 364
Windows Media Player (WMP)
(http://www.microsoft.com/windows/windowsmedia/) 367
Beatnik (http://www.beatnik.com/) 368
Shockwave/Flash
(www.macromedia.com, www.macromedia.com/software/flash/) 369
Who Makes the Salad? Web Designers and Plug-ins 376
Making It Work: Providing Options 377
The “Automagic Redirect” 379
The iron-plated sound console from Hell 381
The Trouble with Plug-ins 381
If Plug-ins Run Free 383
Parting Sermon 384
13 Never Can Say Goodbye 387

Separation Anxiety 387
From Tag Soup to Talk Soup: Mailing Lists and Online Forums 389
A List Apart 390
Astounding Websites 390
The Babble List 390
Dreamless 391
Evolt 391
Metafilter 391
Redcricket 392
Webdesign-l 392
When All Else Fails 392
Eye and Brain Candy: Educational and Inspiring Sites 393
Design, Programming, Content 393
The Big Kahunas 395
Beauty and Inspiration 396
The Independent Content Producer Refuses to Die! 401
Index 403
xiii
00 0732 FM 4/24/01 1:38 PM Page xiv
About the Author

Jeffrey Zeldman has been designing websites since the Crimean War. His personal website at
www.zeldman.com has been visited by millions. Jeffrey is the publisher and creative director of A List
Apart (www.alistapart.com), a weekly magazine “For People Who Make Websites”; cofounder and
leader of the advocacy group, The Web Standards Project (www.webstandards.org); and founder of
Happy Cog (www.happycog.com), a web design agency. He is a featured columnist for publications in-
cluding Adobe Web Center, PDN-Pix Magazine, and Crain’s Creativity Magazine and speaks at web and
design conferences around the world. But what he really wants to do is direct.
xiv
00 0732 FM 4/24/01 1:38 PM Page xv
About the Technical Editor

Steve Champeon is the CTO of hesketh.com, a web services firm in Raleigh, NC, that specializes in dis-
tinctive B2B and corporate sites, vibrant online communities, and high impact applications. He has pro-
vided technical editing on the topics of XML, XHTML, and other web-related topics and was the de-
velopment editor for Jeff Veen’s recent bestseller, The Art and Science of Web Design, published by New
Riders. In addition to his work as an editor, Champeon is a frequent contributor to online and print
magazines for web professionals and is the author of Building Dynamic HTML GUIs (published by IDG
Books Worldwide).
A highly sought-after speaker at trade conferences, Champeon regularly participates in CMP’s Web
conference circuit and Cool Site in a Day competition, Thunder Lizard, South by Southwest (SxSW), and
others, often speaking on DHTML and how to grow successful online communities.
xv
00 0732 FM 4/24/01 1:38 PM Page xvi
Dedication
To Joan, whose love makes me feel happy and safe.
To my Dad, who taught me to be independent.
To my Mom, who loved books. I wish she could have seen this one.
xvi
00 0732 FM 4/24/01 1:38 PM Page xvii
Acknowledgments
I cannot possibly name all the people whose creativity has inspired me, or those I’ve been lucky enough
to collaborate with over the years. It would take hundreds of pages to properly thank those I’ve worked
with this year alone.
In childhood, I attended a wedding where the bride and groom thanked the special people in their lives.
In the flush of the moment, they forgot to name one friend. He harbored a resentment that deepened
over the years. Ultimately, a tragedy ensued, in which innocent bystanders lost their lives. But I digress.
Rather than make a similar mistake, I’m going to deliberately omit the names of many special people
who contributed to my knowledge of the Web and thus, however unwittingly, to this book. Even if you
are not named below, I love you and am grateful to you, and you should buy this book regardless.
To Steve Crozier of Populi, who envisioned an intelligent method of teaching web design, and to Mar-
garet Alston and Cheryl Stockton, who collaborated with me on the development of the Populi Cur-
riculum, my sincere and endless thanks.
My deep gratitude to Michael Nolan for asking me to write this book. To Michael and Karen White-
house for shepherding it safely through the minefields of the publishing industry. To development ed-
itor Victoria Elzey for keeping it real. And to my friend and this book’s technical editor Steven Cham-
peon for finding all the mistakes and not telling anyone but me.
To my beloved friends Fred Gates, Leigh and TJ Baker-Foley, and Katherine Sullivan: thank you for shar-
ing your lives, keeping me sane, and forgiving the disappearances, hibernation, and mood swings that
accompanied the writing of this book.
To Jim, who asked only an occasional phone call and got nothing but months of silence: I wrote this
book for you, I owe you more than these words express, and I promise to start calling again, really.
To Don Buckley, my friend and first web client, and to my first web design partners, Steve McCarron
and Alec Pollak, sincere thanks and respect.
All web designers owe thanks to Glenn Davis for contributions too numerous to describe here. Simi-
larly, respect and thanks to George Olsen, Teresa Martin, and Michael Sweeney. You know what you
did.
Love, thanks, and respect to Brian M. Platz, co-founder of A List Apart back when it was a mailing list
for web designers. To Bruce Livingstone, Nick Finck, Webchick, and Erin Kissane, who help keep ALA go-
ing. And to the fine writers who make it worth reading, including Joe Clark, J. David Eisenberg, Curt
xvii
00 0732 FM 4/24/01 1:38 PM Page xviii
Cloninger, Alan Herrell, Scott Kramer, Jeffrey Veen, John Allsopp, Robin Miller, Denice Warren, Jason
Kottke, Lance Arthur, Glenn Davis, Alyce McPartland, Ryan Holsten, Julia Hayden, Peter-Paul Koch,
Wayne Bremser, D.K. Robinson, L. Michelle Johnson, Mattias Konradsson, Steven Champeon (again),
Chris Schmitt, Marlene Bruce, Lee Moyer, Bob Stein, Dave Linabury, Mark Newhouse, Bob Jacobson, Eri-
ka Meyer, Ross Olson, Rich Robinson, Bill Humphries, Scott Cohen, Peter Balogh, Robert Miller,
Shoshannah L. Forbes, Pär Almqvist, Simon St. Laurent, Jennifer Lindner, Nick Finck (again), Jim Byrne,
Makiko Itoh, Ben Henick, George Olsen (again), and Chris MacGregor.
Thanks to everyone who’s ever looked at any site I’ve had a hand in creating, and especially to those
who’ve written (even if you wrote to say it stank). Thanks to all the web designers and developers who
joined The Web Standards Project.
Hello? Thanks to Tim Berners-Lee for inventing the Web. Thanks to the Web’s first teachers: Jeffrey Veen
(again), Glenn Davis (again), Dan Shafer, David Siegel, and Lynda Weinman. Thanks to Jim Heid and
Steve Broback of Thunder Lizard for support, encouragement, great programs, and fine hotel accom-
modations.
Thanks to Michael Schmidt and Toke Nygaard for the secret work you did on this book, for the incred-
ible work you do on the Web, and for your friendship. Similar thanks to the incredible Carlos Segura.
Thanks to Todd Fahrner and Tantek Çelik for contributing to my knowledge and (more importantly) to
the sane advancement of the Web. Likewise, each in their own way: Tim Bray, Steven Champeon (again),
Rachel Cox, B.K. DeLong, Sally Khudairi, Tom Negrino, Dori Smith, Simon St. Laurent, Eric Meyer, Eric
Costello, J. David Eisenberg (again), Dave Winer, Stewart Butterfield, Carl Malamud, Joe Jenett, Evan
Williams, Robert Scoble, and Peter-Paul Koch (again).
Huge shout-outs to my supremely talented web designer pals. I value your friendship and love your
work. You know who you are, and if you didn’t know you might get a clue from the fact that I am al-
ways linking to you or referring obliquely to you, and if that’s not enough, you’ll find yourselves in the
Exit Gallery at zeldman.com.
The paragraph above and the one you’re now reading constitute the toughest part of writing this book.
In the six years I’ve spent designing websites, I’ve met or corresponded with tens of thousands of tal-
ented people, worked with or gotten close to hundreds. I can’t list you all. This is so painful I feel like
canceling the book, but my publisher insists otherwise. Please accept these tragically empty paragraphs
as my attempt to embrace you all in love and gratitude.
Love and thanks to Peyo Almqvist, Derek Powazek, Josh Davis, Heather Champ, Daniel Bogan, Craig
Hockenberry, Lance Arthur, Michael Cina, Heather Hesketh, Dave Linabury, Dan Licht, Brian Alvey, Shau-
na Wright, Halcyon, Hasan, Matt, Jason, Big Dave, Lmichelle, Fish Sauce, Toke, Michael, Leigh, and
Uncle Joe.
xviii
00 0732 FM 4/24/01 1:38 PM Page xix
Foreword
I wrote this book for four people:
For Jim, a print designer who’s tired of sending his clients to someone else when they need a website.
For Sandi, a gifted art director, who’s hit a wall in her advertising career, and is eager to move into full-
time interactive design.
For Billy, whose spare-time personal site has gotten so good, he’d like to become a professional web
designer—but is unsure about what is expected or how to proceed.
And for Caroline, a professional web designer who wants to better understand how the medium works
and where it is going.
I did not make up these names or descriptions: These are real people. I knew the book was finished
when it had covered everything they needed to know.
An entire curriculum, a year of work and thought, and 100 years of professional experience (mine, my
editors’, and my collaborators’) have gone into this book.
Enjoy.
Jeffrey Zeldman
1 April 2001
New York City
xix
00 0732 FM 4/24/01 1:38 PM Page xx
Web vs. Print: A Note About URLs

The Web is an ever-changing flow of ideas, designs, and redesigns. Sites evolve and decay. Some move
to new locations. Others disappear. By the time you read this book, some of the sites it describes will
surely have changed, while others may have vanished altogether.
This flow and flux is natural to the Web, and in some ways it is even healthy. It’s good when mediocre
sites improve, and it’s inevitable that pointless sites (like pointless products) eventually fade away.
But healthy and natural or not, the medium’s constant dynamism can wreak havoc with books about
the Web, and thus with those books’ readers. You read about an interesting design or technological de-
cision, fire up your web browser, and discover that the site no longer demonstrates what was discussed
in the book.
Fortunately, dear reader, you can minimize the damage by bearing these things in mind:
1. Most of the concepts and techniques discussed here are fairly widespread. If Site A no longer
sports a nifty rollover technique we’ve described, you’ll probably find it at Site B or Site C. The
principles are more important than the specific examples.
2. Sites should not arbitrarily change page locations, but unfortunately, many do. If a particular web
page seems to have disappeared, try factoring the URL to a simpler version. For instance, if
www.yahoo.com/games/thrills/ no longer works, go back to its purest form, www.yahoo.com/,
and see if you can navigate to the page’s new location that way.
3. Finally, if a site we’ve hailed as an example of creative excellence or touted as a superb resource
for further learning seems to have disappeared, try visiting the zeldman.com Exit Gallery at
www.zeldman.com/exit.html. If the site is truly special and has moved to a new location on the
Web, you’ll find that new address in our Exit Gallery. If the site has actually changed its name,
we’ll mention the former name to help you get your bearings.
Now go forth, design, and conquer.
xx
01 0732 Intro 4/24/01 11:12 AM Page 1
Introduction
WHEN WE FIRST MET STEVE CROZIER, president of Populi, we liked what he had
to say.
He said, “I want to buy you lunch.”
When he told us his company’s vision, we liked that even more.
It was a simple solution to a complex problem. On one side, thousands of

designers and art directors are eager to take their talents to the Web but
aren’t sure how. On the other, web agencies could not find enough good
web designers to get their work done.
The Populi program was designed to close the web talent gap by training
traditional designers in the ways of the Web. Until ithe Populi program
comes to your town, this little book can teach you what you need to know.
This is not one of your “Learn HTML in 24 Hours” books, nor is it one of the
many introductory books on web graphics. It won’t teach you how to imi-
tate the stylistic tricks of famous web designers, turn ugly typography into
ugly 3-D typography, or build online shopping carts by bouncing databases
from one cryptic programming environment to another. This is a book for
working designers who seek to understand the Web as a medium and learn
how they can move to a career in web design. It’s also suited to designers
who wish to add web design to their repertoire of client services.
01 0732 Intro 4/24/01 11:12 AM Page 2
2 Introduction
Populi (www.populi.com),
the Web Talent Incubator,
turns traditional designers
and programmers into web
builders.
Why did we base this book on the Populi curriculum? For one thing, it’s one
of the only programs we know that actually works. For another, we wrote
the curriculum. (To be honest, we wrote the curriculum in cooperation with
courseware developer Margaret Alston, and designer-instructor Cheryl
Stockton, of the Pratt Institute. The cranky opinions are ours; the thor-
oughness and good sense—theirs.)
The concepts contained in the Populi curriculum and this book have been
field-tested on working designers. They’ve been reviewed by web agency
consultants and Pratt faculty members, spoken aloud to tens of thousands
of web conference attendees, rolled in flour, and slow-baked at 450
degrees.
This book will teach you how web design compares to and differs from the
job you know and love. It will explain the medium’s challenges, such as
bandwidth, navigation, and browser compatibility. And it will teach you
enough of the technical details to work with your peers on the production
end and to pinch-hit as needed.
The Populi Curriculum in Web Communication Design, created in coopera-

tion with Pratt Institute, was launched in Dallas in 2000 and will eventu-
ally come to your town.
On the other hand, the book you are holding is available now, at a modest
price.
You know what to do.
02 0732 Part I 4/24/01 11:13 AM Page 3
Part I
WHY: Understanding the Web
1 Splash Screen 5
2 Designing for the Medium 13
3 Where Am I? Navigation & Interface 69
02 0732 Part I 4/24/01 11:13 AM Page 4
03 0732 CH01 4/24/01 11:14 AM Page 5
chapter 1
Splash Screen
WHAT DO DESIGNERS DO? Designers organize information, shape identities,

and create memorable experiences that entertain while communicating.
Increasingly, designers are performing these tasks on the World Wide Web
(the Web, to its friends). If you’ve picked up this book, you’re either doing
the work already, thinking of migrating to the field, or considering adding
web design to your repertoire of existing services.
Whether you design websites full-time or just occasionally, you’ll be help-

ing to shape what may be the most inherently profound medium since the
printing press. The Web is vast, intrinsically democratic, and dripping with
creative, personal, and business potential. Oddly enough, for something
that gets used and talked about every day by hundreds of millions, it is also
quite often misunderstood by practitioners as well as users.
Before you do anything drastic, such as buying “web software,” changing

your career, or leaving that louse who is only pretending to love you, it
makes sense to find out where you are going and what you will be dealing
with. So let’s start by examining what the Web is—exactly.
03 0732 CH01 4/24/01 11:14 AM Page 6
6 WHY: Splash Screen: Meet the Medium
MEET THE MEDIUM

The Web is a part of the Internet, a group of interconnected computer net-
works that spans the globe. Web servers deliver content of many kinds,
much of it connected to other content via hyperlinks and therefore
referred to as hypertext. Most of these documents are written in a simple
markup language called HTML, about which we will have much more to
say. But web servers aren’t limited to publishing HTML documents; they can
deliver almost any digital content you care to envision.
Put another way, the Web is a medium, like print or television. It is avail-
able worldwide to anyone with an Internet connection. Unlike with print
or television, though, the Web is a two-way street. Not only can anyone
with an Internet connection view and interact with websites, he or she also
can create or contribute to such sites.
At this moment in history, the Web is usually experienced on a desktop

computer. This is changing rapidly, though, as web-enabled cell phones and
Palm Pilots become Yuppie accessories that make you just want to slap
them. (The Yuppies, not the accessories.)
Desktop web browsers, such as Netscape Navigator, Microsoft Internet

Explorer, and Opera Software’s Opera, are used to view and interact with
the content on websites. These “sites” are collections of web documents
published online at specific virtual locations. They’re called sites, not books,
because the Web is not print and because the founders of the Web were
obsessed with solving basic problems such as that of location. Where do
web documents go? Where can people be assured of retrieving them? The
founders of the Web developed a system of Uniform Resource Locators
(URLs), affording every web document the luxury of a permanent address—
hence, a site collection, not a book collection.
By the way, while URLs make possible a permanent address for every web
document, such permanence is not guaranteed. Companies go out of
business and take down their sites; products are replaced by newer mod-
els, and the old web pages go offline; news and information sites hampered
by limited server space kill old stories to make room for new ones; or a
03 0732 CH01 4/24/01 11:14 AM Page 7
Taking Your Talent to the Web 7
new publishing system comes online, and old web addresses such
as www.url.com/issues/01/03/story.html are replaced by new robot-
generated URLs such as www.url.com/content.cgi?date=2001-03-21/
article.cgi?id=46&page=1.
Outside the corporate web sphere, personal sites go offline when their cre-
ators get bored or they move to a new location, and the creator neglects
to leave a forwarding address. There are as many scenarios as there are web
pages that have disappeared. This is a problem for web users who book-
mark certain pages in hopes of revisiting them and for directories such as
Yahoo.com or search engines such as Google.com whose business is to con-
nect seekers of specific information with sites that meet their needs.
Expanding Horizons
Searches and similar activities underscore the fact that the web experience
is interactive—another difference between it and print and TV. Visitors not
only link from page to page at their discretion, they also can post their own
content to some sites, shop at others, play games, or alter the design ele-
ments to suit their tastes at still others.
Needless to say, these interactive aspects of the Web present incredible

design challenges and opportunities, which grow more interesting and
more sophisticated as the Web’s capabilities expand. And they are expand-
ing every minute. While we wrote this book, Microsoft came out with IE5.5,
Opera unveiled Opera 5, Netscape produced Navigator 6, and Macromedia
premiered Flash 5. To varying degrees, all four products have changed pro-
foundly what the Web can be—the three browsers by offering increased
support for powerful web standards such as CSS, XML, and the DOM and
Flash 5 by providing richer (though proprietary) design and programming
tools.
Note
We will discuss CSS, XML, and the DOM in due course. If you're nervous or
simply curious, skip ahead to Chapter 5, "The Obligatory Glossary," then come
on back.
03 0732 CH01 4/24/01 11:14 AM Page 8
In terms of technological acceptance, the Web has grown faster than any
medium in history. In 1990, there were two “wired persons” (people con-
nected to the Web): Tim Berners-Lee, the physicist who invented the Web,
and his friend and colleague Robert Caillou. By 1993, there were 90,000
web users. Two years later, there were three million. By early 1999, that
number had grown to over 200 million, 80 million of them in the U.S. alone.
Six months later, estimates were well over 300 million. Soon there will be
more web users than McDonald’s burgers sold. Fortunately, no animals
were harmed in the making of the Web.
Computers will always be unaffordable for some folks, and others simply
dislike technology. How will the Web keep growing after everyone with the
means and desire has bought a computer and a modem (or whatever high-
speed connectivity that replaces the modem)?
It will grow by slipping past its existing borders. Drivers will receive direc-
tions from devices in their cars without realizing that the data is stored on
a site you may have designed. Technophobes will interact with sites while
finding out local movie times over the phone. They won’t know they’re get-
ting information from the Web; for them this will simply be a conventional
telephone experience. You won’t be responsible for porting the data (geek-
speak for translating web content into something a web-enabled phone
can understand), but your sites will undoubtedly reach people who have
never touched a “traditional” web browser.
Within the next five to ten years, it’s fair to say that “everyone” will use the
Web, just as “everyone” uses the telephone. Of course, there are human
beings who don’t use the phone (and many who don’t answer it, especially
if they owe you money), but we’re speaking in generalities to emphasize a
simple point:
You are about to begin designing for a medium that will eventually
reach practically every home and office in every corner of the world.
Your work will potentially affect the lives of billions. You will never
be lonely or go hungry again. But on the flip side of that joyous
news, you will face new challenges and will need to learn new skills
throughout your web career.
03 0732 CH01 4/24/01 11:14 AM Page 9
“Billions” sounds like a pretty daunting audience. But as with all design,
remember that you’re not trying to reach or please everyone. If you design
to communicate ideas and if your clients are focused enough to have prod-
ucts or causes worth sharing with specific people, then the right hundreds,
thousands, or millions of people will visit and be enriched by your sites.
“Your sites.” It sounds nice, doesn’t it?
Working the Net…Without a Net

Given this vast, worldwide audience, you will no longer be able to assume
certain things—for instance, that everyone who visits the site speaks Eng-
lish. Or that every visitor has an equally powerful computer, an equally up-
to-date browser, or an equally glorious monitor with which to view your
work. You can’t even assume that all your visitors can view your work at
all, in the conventional sense of that word. Millions of people with visual
(and other) disabilities use the Web every day; believe it or not, your
designs can accommodate them. (We’ll talk about how that’s done
throughout the book.)
In art direction and graphic design, before you even begin conceptualizing
your approach, you must target your audience and learn the size of the
medium you’ll be working with (magazine spread, quarter-page newspa-
per, or outdoor billboard). On the Web, audience projection is an imperfect
science at best, and there are no absolute sizes, or absolute anything else.
But don’t reach for the Absolut vodka—there’s nothing to fear. Your design
vocabulary is simply going to enlarge. In fact, your whole conception of
what it means to design will expand.
While it broadens in its reach, the Web also is constantly increasing

its capabilities—from the early, text-only Web, to text plus images, to
streaming media (audio, video, and multimedia environments created in
Flash, Shockwave, and Java). From static pages, to dynamically generated
pages, to sites to which the word “page” does not apply at all. (For a
taste, visit www.eneri.net, www.photomontage.com, or www.once-upon-
a-forest.com.)
03 0732 CH01 4/24/01 11:14 AM Page 10
Most of the time in this book, we’ll be discussing the Web as we know it
and as your clients understand it: an interactive digital medium accessed
via a desktop computer with an Internet connection and viewed by means
of a web browser such as Internet Explorer, Netscape Navigator, or Opera.
We also will assume that you’ve used the Web yourself. Maybe you while
away the hours in a chat room where you’re known as HotBuns32 or you
spend half your life checking other people’s bleeding edge site designs. Per-
haps you just check your email once a week (and pretend you haven’t read
it when it’s from a relative) or log on once a year to save $5 on a Mother’s
Day bouquet. If you haven’t done any of this, go online now, and we’ll talk
later.
Though we’ll focus on the Web you know, we also will talk about the ways
the Web is changing—because those changes will have as profound an
impact on your career as they will on our civilization. What you’ll learn in
this book is only the beginning. (If you’re not comfortable with the idea
that a career in web design necessitates continual learning, put this book
down now and back slowly away.)
On the other hand, you might like the idea that the Web is steadily expand-
ing its borders. That people can already access some web content via hand-
held devices such as Palm Pilots. That there are web phones out there and
browsers for the blind. That web-based navigation systems are finding their
way into the cars and trucks we drive. That there is actually a prototype
web refrigerator, and that before we get much older almost every device
imaginable will be accessing the Web in some way or other—whether it
needs to or not.
All these applications will require the skills of talented designers (and pro-
grammers, of course). So congratulations on making an absolutely brilliant
career move. Now buy this book so you can actually start doing something
about all this.
If you’re curious about how the heck this Web thing got started, see Chap-
ter 4, “How This Web Thing Got Started.” If you’re unsure of your termi-
nology, see Chapter 5, “The Obligatory Glossary.” You’ll find both chapters
in Part 2, “WHO,” in the middle of this book, along with useful material on
03 0732 CH01 4/24/01 11:14 AM Page 11
the project life cycle and a detailed definition of the web designer’s role. If
you’d like to hear more about how smart you are for deciding to learn about
web design, phone your Mom—that is, if she’s forgiven you for that cheap
floral bouquet you got her.
On the other hand, if you’re ready to plunge into the most interesting
aspects of web design, Chapter 2, “Designing for the Medium,” has your
name on it, baby. But before you dive into it, we need to make one more
prefatory point.
SMASH YOUR ALTARS

With the exception of a few facts, everything in this book is subject to
debate. Web design, like the medium, is too new to be bound by fusty rules.
When we explain general principles and accepted practices, our goals are
to clarify how the medium functions and to ground you in the thinking and
methods of most working web practitioners. You will need to know this in
order to do your job. But it is only the beginning, and you are encouraged
to constantly think beyond everything we tell you here.
For every ten sites that fail because they’ve ignored a certain web verity
(for instance, that navigation should be clear and streamlined), there is at
least one site that succeeds precisely because it violates this “rule” in a
unique and brilliant way. For every hundred sites that fundamentally mis-
understand the medium by behaving like static Illustrator layouts, there is
one that achieves greatness by doing so.
Most web designers begin each project by considering the end-user. But
we know of at least one certifiable web design genius who starts every job
by inventing dynamic behaviors he has never seen on anyone else’s sites
and then following those behaviors wherever they lead. Remarkably
enough, they lead to professional and usable sites whose uniqueness
delights precisely the users they were intended to serve. This should not
work at all, but it not only does work, it enlarges what the Web can be.
There is stupidity (and there is a lot of it). And then there is innovation and
creative rule-breaking that sometimes leads to greatness.
03 0732 CH01 4/24/01 11:14 AM Page 12
12 WHY: Splash Screen: Smash Your Altars
If your boss or client dictates or forbids a certain web design practice

because of some rule in an old web book (or, sadly, in a new book full of
bad ideas), we won’t mind you citing this book to counter the argument.
But please don’t invoke this book as an authoritative set of web design
commandments. This is not a book of rules, and any web book that pre-
tends to be is full of it. Take what we say seriously but stay flexible. Musi-
cians learn scales before writing melodies. These are the scales; you’ll write
the tunes.
04 0732 CH02 4/24/01 11:14 AM Page 13
chapter 2
Designing for the

Medium
THE WEB IS LIKE EVERY OTHER MEDIUM to which you’ve applied your talents and
like no other medium you’ve ever grappled with. Everything you know as a
designer will help you tremendously, yet nearly everything you know must
be rethought. Sounds like a sales pitch—until you’ve actually tried your
hand at web design.
The Web is different because websites must function as both documents

and databases. It’s different again because the medium is somewhat
ephemeral in nature, never looking or functioning exactly the same way for
each person who encounters it. Prove this to yourself by visiting any
sophisticated site using IE5 on an iMac, Netscape 6 in Linux, and IE4 on a
Windows PC. If it looks and works exactly the same in all three settings,
we’ll eat our Aunt Miriam’s crepe de chine hat. And these are just three of
thousands of possible combinations.
The Web is both more and less capable than print. On the one hand, it pro-
vides near-instant access to information, offers rich multimedia experi-
ences, and responds dynamically to the visitor’s actions. On the other, it
defeats the designer’s desire to completely control the visual experience.
The Web is different because anybody can make a website, but not every-
body can do it well.
04 0732 CH02 4/24/01 11:14 AM Page 14
14 WHY: Designing for the Medium: Breath Mint? Or Candy Mint?
Finally, the Web is different because it works best when it’s lean and mean.
Looking at a full-bleed, two-page spread places no strain on magazine
readers, but viewing sites that make extensive use of images, sounds, and
other “heavy” media can put a serious crimp in the web user’s experience—
particularly if the designer has not taken pains to optimize the site. File
sizes must be kept small if web pages are to download quickly and effi-
ciently over slow, dialup modem connections (or even fast connections).
Include too many images or other files per page, and the fastest connec-
tion will slow to a crawl due to limitations in the number of files that can
be served simultaneously.
This conflict between size and speed is known as bandwidth, and we will
have much to say about it later in this chapter. For now, the following dis-
turbingly technical definition will either give you your bearings or send you
screaming back to the safety of print design.
A Definition of Bandwidth
According to Whatis.com (www.whatis.com):
"Bandwidth (the width of a band of electromagnetic frequencies) is used

to mean… how fast data flows on a given transmission path…. It takes more
bandwidth to download a photograph in one second than it takes to down-
load a page of text in one second. Large sound files, computer programs,
and animated videos require still more bandwidth for acceptable system
performance."
Designing for the medium is a joy—once you understand the Web’s limita-
tions and opportunities.
BREATH MINT? OR CANDY MINT?

If you know your web history (or if you’ve skipped ahead to Chapter 5, “The
Obligatory Glossary”), you’ll recall that the Web was conceived as an open
platform for distributing structured text documents. When physicists Tim
Berners-Lee and Robert Caillou created Hypertext Markup Language
(HTML) as a limited subset of a much more complex open standard for doc-
ument publishing, graphic design was the last thing on their minds.
04 0732 CH02 4/24/01 11:14 AM Page 15
HTML was as simple as rain. It was built in that way so scientists could learn
it quickly and use it to publish their physics papers online. Documents pub-
lished in HTML were “styled” by the default settings of early Web browsers
(the familiar Times New Roman on a gray background). Early web pages
looked exactly like physics papers, which was pretty darned great if you
were a physicist.
But clients don’t buy physics papers. After designers and their clients
grasped the Web’s commercial potential, they began seeking ways to make
web pages look as good as other professional publications. Today, web
standards such as Cascading Style Sheets (CSS) allow us to do just that.
But in 1994 and 1995, these standards did not exist, so web designers and
browser makers such as Netscape began “extending” the behavior of HTML
in nonstandard ways.
What happened to HTML was not unlike what happens to legislation intro-
duced in the U.S. Congress. A legislator wants to change the speed limit in
his home state. By the time it gets out of committee, the bill includes taxes
on liquor and tobacco, gun licensing restrictions, subsidies for farmers,
mandatory parental warnings on CDs and cassettes, and an impassioned
plea for school prayer. Over the years, HTML was similarly amended,
extended, and tacked onto by a thousand hands. Many of those amend-
ments were intended to facilitate the needs of designers. A few were just
plain wacky. We’ve been coping with the damage ever since. Take the fol-
lowing example:
HTML in the “Good Old Days”:
<a href=”index.html”><IMG SRC=”image.gif” alt=”Return to the home page.”></a>
HTML Today:
<tr><td valign=”top”><a HREF=”index.html” target=”elchico” onMouseOver=”

window.status=’Home again, home again, jiggity jig.’; changeImages(‘toc’, ‘omen2/
coreover.gif’); return true;” onMouseOut=”window.status=’’; changeImages(‘toc’,
‘omen2/core.gif’); return true;”><img name=”toc” src=”omen2/core.gif” width=”49”
height=”25” border=”0” alt=”Return to the core page.” Title=”Home again, home again,
jiggity jig.”></a></td></tr>
04 0732 CH02 4/24/01 11:14 AM Page 16
Later in this chapter, we’ll talk about HTML and web standards in more
detail. For now, it’s important to realize that the impulse behind the Web’s
creation was logical, structured, and intended to address a basic need: the
simple sharing of data. It was never about marketing or design.
Despite all that has befallen since those early days, many people continue
to view the Web as an archive or database of searchable information. And
some of these folks have espoused a set of “rules” to ensure that web pages
yield their information with a minimum of fuss and confusion. Let’s call this
group the Usability People. Jakob Nielsen is one of their foremost expo-
nents, and you can read what he has to say at www.useit.com. To Usabil-
ity fans, anything that impedes access to the data is bad; anything that
momentarily confuses even a single user is bad; and thus, pretty much any-
thing out of the ordinary is viewed with suspicion or banned outright. This
view of the Web is straightforward and can serve as a touchstone for web
designers, though the guidelines espoused by Usability gurus should not be
confused with Commandments. (Last time we checked, the Command-
ments were written by Someone else.)
Usability basically reminds designers to think about the needs of their

audience. On many commercial and informational sites, web users simply
hope to find things or do things as quickly as possible. When checking
sports scores or seeking low airfares, they do not wish to be creatively chal-
lenged by a complex multimedia experience. They merely want to find what
they seek and get on with their lives.
This does not mean that web design is a cold, calculating science. Far from
it: Like all good design, web design is aesthetic, emotional, and largely
unquantifiable. The value in the Usability perspective is that it reminds web
designers to create sites that people can actually use.
This ought to go without saying, but you’ll find that in web design almost
nothing goes without saying. Perhaps in print you’ve known designers who
become so carried away with graphic design for its own sake that they for-
get to communicate. The synergy between headline and visual gets lost in
a haze of technique; typography advances toward illegibility; subtleties of
lighting completely obscure the subject, and so on. When web designers
make the same mistake, potential readers and customers are thwarted in
04 0732 CH02 4/24/01 11:14 AM Page 17
their desire to use the site. The folks in suits start beating the designers
over the head with Jakob Nielsen’s latest book, and a good time is had by
no one. Don’t let this happen to you. It’s easy to avoid if you keep the
intended user and usages in mind.
Magazine and ad layouts may be wild or restrained as long as they are leg-
ible. Web design must be much more than legible, though many sites fail
to achieve even basic readability, and few indeed are a pleasure to read. (To
say nothing of the fact that most ad layouts are intended to convey sim-
ple messages, while websites often perform numerous, complex functions.)
In his widely read 1996 treatise, Creating Killer Websites, David Siegel
listed three cardinal virtues of web design: “Clarity, Brevity, Bandwidth.”
Though Siegel was a graphic designer and not a Usability Person (and
though he did not always achieve these goals in his own work), there’s
likely not a Usability Person on the planet who would disapprove of that
trinity.
But many designers and artists saw something quite different in the Web:
the chance to create and publish creative works that plunge the viewer into
a unique world of imagery, exploration, and cinematic or personal narra-
tive. This view, implicit in sites such as Photomontage (www.photomon-
tage.com) and Presstube (www.presstube.com), is as vital to the health of
the medium as the contrasting Usability perspective. We’ll call its expo-
nents the web artists, though this label is somewhat misleading. For while
it’s true that many web artists are motivated by the urge toward pure cre-
ative expression, the trails they blaze are invariably followed by marketers
in search of deep online branding opportunities. The innovations delivered
by pioneering multimedia artists quickly become the basis for sites touting
Motown, Madonna, or Barney’s New York.
Web artists do not believe in holding the visitor’s hand. They judge that
websites can be as challenging as paintings, music, literature, or Swedish
movies. They further hold that there is an audience for sites that raise bars
and test boundaries. They are, of course, correct. Challenging sites can
reward patient viewers. They don’t please everyone but neither does mod-
ern painting. Writer Curt Cloninger summed up the conflict between those
who view the Web as an informational database and those who see it as
a wide-open aesthetic frontier when he shrugged, “Usability Experts are
from Mars, Graphic Designers are from Venus” (www.alistapart.com/
stories/marsvenus/).
04 0732 CH02 4/24/01 11:15 AM Page 18
Figure 2.1
Supermodified looks like
(and is) a work of multi-
media art. Yet it serves a
commercial purpose.
Visitors can trigger loops
of music by typing on the
keyboard. A strictly infor-
mational approach to site
design, such as the Google
Search Engine (Figure 2.2),
would be far less effective
at creating excitement
about the composer’s work
(www.amontobin.com).
Figure 2.2
The Google Search Engine.
A classic example of func-
tion driving form (with the
possible exception of the
logo). Google’s search
engine delivers solid
results, and hardcore geeks
love it because it strips
away the clots that clog
the arteries of most com-
mercial search engines.
Both Google and amonto-
bin.com are successful at
doing what they set out to
do, yet they are clearly
different in their approach
to the user experience
(www.google.com).
04 0732 CH02 4/24/01 11:15 AM Page 19
Mars and Venus, left and right brain, utility and artistry. On one side stands
a set of Usability Commandments based on roughly a decade of trial
and error and a heaping teaspoon of pseudo-science. On the other lies the
indefinable essence of art and a horde of marketers who stand ready to
exploit it.
Somewhere between these two extremes you will find the appropriate bal-
ance for each site. The ideal balance for most sites will not be found in the
stone tablets of Mars or the sensual abandon of Venus. Rather, it will come
from each project’s intended audience. Your visitors’ needs set the param-
eters; your taste, inspiration, and expertise do the rest.
That tension between structure and style, function and aesthetics, is key to
understanding web design and web technology. Users have needs; tech-
nology sets limitations. The conflict will resurface throughout this book and
your career—and it is only the beginning. Web design is different in fasci-
nating ways. Following are a few key points of difference.
Where’s the Map?

Books, magazines, CDs, and videocassettes do not need to explain them-
selves. Most of us read from left to right and top to bottom; we turn the
page. We insert the disc or tape and press Play. Websites are not so self-
explanatory. Consequently, web designers spend a great deal of effort cre-
ating contextual and navigational cues to guide readers, viewers, and
“users” through the site.
Visitors take their cues from non-web experiences. From a lifetime of

newspaper reading, they know that headlines carry more weight than sub-
heads and body copy. They intuitively grasp that right-pointing arrows
mean “more” or “continue.” (This intuitive grasp is, of course, the result of
previously absorbed social conventions. Red, green, and yellow buttons
suggest traffic lights to an American web user; they may mean something
different or nothing at all in Papua, New Guinea.) Web users also take their
cues from other sites they’ve seen. Soon after figuring out how the modem
works, users learn that underlined text is almost always a link, and they
know that when the cursor changes shape they are hovering over an
“active” link or image.
04 0732 CH02 4/24/01 11:15 AM Page 20
20 WHY: Designing for the Medium: Web Physics
Mars and Venus

Adept web designers take care to follow some familiar contextual conven-
tions while breaking or reinventing others. On one site you might use CSS
to turn off link underlining; on another, you preserve link underlining
because the site is intended for neophyte users who need to be led by the
hand. One site requires idiot-proof icons with text labels; another cries out
for subtle, dynamic navigational menus. Usability People lecture sternly
about the “sins” of web design, but designers don’t sin—they make deci-
sions. A good web designer may break as many rules as she follows. Visi-
tors determine whether the site succeeds as a piece of communication or
is merely a failed, cryptic experiment. This book explores issues of naviga-
tion and interface in Chapter 3, “Where Am I? Navigation & Interface.”
You’ll be exploring them for the rest of your career.
That we devote an entire chapter to navigation and interface should be

indication enough that graphic design alone does not equal web design—
a point we’ll restate several times in case some of you haven’t had your
coffee yet. Choosing and setting type, crafting pretty buttons, and devel-
oping a grid are all well and good but not good enough. Above all, web
designers are the architects of user experience.
You might feel that your training and experience have not prepared you to
build such architecture, but you’ll soon see that it’s the web equivalent of
what a designer always does: guide viewers toward an understanding.
WEB PHYSICS: ACTION AND INTERACTION

Design for the Web is different. It’s different because web pages don’t just
sit there; they do things. More importantly, they allow visitors to do things.
Magazine pages may be beautiful (or not) but the reader’s interactivity
consists of reading the page (or not), dog-earing it (or not), and rereading
it (or not). At most, the reader might cut it out and mail it to a friend.
Strictly speaking, none of this can truly be called interactivity. Beautiful
magazine layouts do not change in response to the viewer’s actions. News-
paper ads do not sprout additional body copy if the reader shows genuine
interest. The Web invites depth of exploration in ways traditional media
cannot. For a designer, the creative possibilities are tantalizing and practi-
cally limitless.
04 0732 CH02 4/24/01 11:15 AM Page 21
On the Web, linear motion gives way to user emotion. Site visitors link ran-
domly as they choose. Set up as many careful hierarchies and navigational
cues as you want; visitors will still do what they like on most sites. Not only
may visitors move up, down, and sideways, they also can bookmark any
page they fancy; download it to their hard drives; save the images from it;
and even study the HTML markup with which it was produced.
Readers can order books on the Web by typing in HTML form fields sup-
ported by scripts written in Perl, Java, or other programming languages
(www.amazon.com). They can post their opinions to message boards
(www.metafilter.com). If the designer has given them the option, they may
change the background colors to suit their mood (www.camworld.com). On
fancy Dynamic HTML (DHTML) sites, they can drag images from place to
place (www.dhtml-guis.com/game/). On fancier ones, they can do much
more (www.assembler.org). On a corporate intranet site, employees may
spend hours updating a group calendar or adding phone numbers to a con-
tact database. (Anything to avoid working.)
Figure 2.3
Non-commercial interac-
tivity: Assembler.org was
created with DHTML (here
it is done well). As of this
writing, the site was opti-
mized for Netscape and
Microsoft’s 4.0 browsers,
which rely on proprietary
coding techniques. Thus
the site’s marvels would
be invisible to users of
recent browsers that avoid
proprietary, old-school
DHTML. By the time you
buy this book, the site
should function well in
standards-compliant
browsers such as Netscape
6 (www.assembler.org).
04 0732 CH02 4/24/01 11:15 AM Page 22
22 WHY: Designing for the Medium: Web Physics
Figure 2.4
Commercial interactivity:
Barnes & Noble, a func-
tional and attractive
shopping site. Successful
e-commerce sites work in
as many browsers as pos-
sible and add value to the
commercial transaction by
providing content and
artificially intelligent
“shopping tips.” Though
Barnes & Noble has a
real-world heritage,
Amazon.com dominates
the online market because
Amazon came first. When
web brands are effective,
users can be incredibly
loyal (www.bn.com).
There is obvious commercial value to commercial interactivity; novelty or
“proof of concept” value to dynamic artwork and games; branding value to
interactive multimedia (www.barneys.com); and hidden value to still other
types of interactivity. (Changing the background color may seem trivial to
you or me, but it could be vitally important for a color-blind web user.)
Overall, interactivity is a defining characteristic of the Web and thus of web
design. Lesbian poetry and physics papers did not drive the rapid expansion
of the medium. Commerce did that, and commerce depends on interactiv-
ity: the visitor clicks, the sale is made.
No offense to the lesbian poetry sites. In fact, no offense to the hundreds

of thousands of noncommercial sites that bring richness, depth, and mean-
ing to the Web. Without these noncommercial sites, the medium would be
nothing more than a dialup variation on the infomercial. But without all
the commercial sites, the Web’s infrastructure, services, and rate of adop-
tion might never have grown so quickly.
At least, that’s what the marketers tell us. Consider this another
Mars/Venus variation for your pleasure. The Internet grew in popularity for
at least two years before any commercial sites were allowed on the Net,
04 0732 CH02 4/24/01 11:15 AM Page 23
much less the Web. And many defining characteristics of the modern Web
($20 unlimited access dialups, 56K modems, free browsers) were estab-
lished by 1995-96, a time when most web users were also web designers,
and the word “commerce” did not begin with the letter “e.” Still, the Web
has expanded like nobody’s business since business came online. And if you
ask most normal humans who’ve gone online in the past few years why
they bought a computer and signed up for an Internet account, “shopping”
seems to top the list.
Different Purposes, Different Methodologies

It is still possible for a lone web designer or small team to create personal,
artistic, and corporate sites using an image editor, HTML, style sheets, and
JavaScript. But the “lone rider” approach is increasingly rare in the corpo-
rate web development space. Today, teams of specialists with odd-
sounding job titles develop most sites collaboratively. (See Chapter 5, "The
Obligatory Glossary” and Chapter 7, “Riding the Project Life Cycle,” for the
funky titles and the typical web project life cycle.) It is not your job to pro-
gram a shopping cart or develop a database. It is your job to understand
where your work fits into the bigger picture.
As a professional web designer, you will work closely with programmers to

implement the appropriate interactivity in every site. You also might be
called upon to execute rudimentary interactivity yourself—for instance,
writing JavaScript to swap images on navigational menus.
WEB AGNOSTICISM
Design for the Web is also different because the Web is not a fixed medium.
It has no size, no inks, no paper stock. Even your typographic choices may
end up as mere suggestions. That’s because the Web is platform-agnostic
and device-independent. Good web design adapts to different browsers,
monitors, and computing systems. What’s sauce for the goose may not be
sauce for the gander. More literally, what’s Geneva for the Mac may be Arial
for Windows; what’s VBScript for Windows may be error messages for Mac
and Linux users. (So don’t use VBScript to build websites.)
04 0732 CH02 4/24/01 11:15 AM Page 24
24 WHY: Designing for the Medium: Web Agnosticism
Looking at poorly implemented sites, you could come away with the
impression that the Web is a Windows application or even an extension of
the Windows desktop. And there are certainly marketers who’d like you to
believe that. But it just ain’t so.
Berners-Lee and Caillou invented the Web on a NeXT computer. The first
browser ever released was for UNIX, the second for Mac OS. Berners-Lee
envisioned the Web as a completely portable medium—one that could be
accessed not only by every computer operating system (including dumb
terminals), but also by all kinds of devices from hand-held Personal Digital
Assistants (PDAs) to telephones and other common appliances. Slowly and
sometimes painfully, everything Berners-Lee envisioned in 1990 has been
coming true.
To help the Web evolve in an orderly fashion, Berners-Lee founded the

World Wide Web Consortium, or W3C (www.w3.org). It’s a place where uni-
versity professors join engineers from companies such as Sun, Microsoft,
AOL/Netscape, IBM, Compaq, and Apple to hammer out common techno-
logical standards, such as HTML and CSS…and more recently, Extensible
Markup Language (XML) and the Document Object Model (DOM). For a
complete listing of W3C member organizations, see the following web
page: www.w3.org/Consortium/Member/List.
Don’t worry about what the acronyms stand for at the moment. Just dig
the concept: If everyone supports the same standards (or “Recommenda-
tions,” in W3C parlance), then designers and programmers will have the
tools they need to deliver a dynamic and attractive Web that works for any
human being, on any platform or device. Sweet, smart, simple.
Sadly, due to competitive pressures, the desire to innovate, and sheer

cussedness, the companies that make web browsers have not always done
a superb job of implementing commonly shared standards. In fact, until
quite recently, you could argue that their support for these standards was
sometimes downright shoddy. You might even be forgiven for suspecting
that browser makers deliberately avoided fully implementing any standard
for fear that supporting common standards would hurt business.
04 0732 CH02 4/24/01 11:15 AM Page 25
In the beginning of this chapter, we mentioned that the Web was spawned
as a beautiful medium for the delivery of physics papers. And that to deliver
commercially viable sites—sites with some semblance of visual appeal—
web designers felt they had no choice but to “hack” HTML, forcing the
deliberately primitive markup language to serve their aesthetic needs.
Netscape (now AOL) joined web designers in extending HTML beyond its
creators’ intentions.
Initially, the Web was a one-horse town. If you wanted to design a com-
mercial site, you wrote nonstandard HTML that was “optimized” for
Netscape’s browser. Once Microsoft’s browser entered the picture, all hell
broke loose, as two powerful software companies began deforming HTML
in mutually exclusive ways.
Browser development was originally viewed as just another genre of soft-

ware development. Adobe Illustrator competes with Macromedia Freehand
by offering features Freehand lacks. Freehand does the same to Illustrator.
God Bless America.
Similarly, Netscape competed with Microsoft (and vice versa) by offering

functionality not supported by the competitor’s product. Each company
hoped these unique features would seduce web developers into creating
sites optimized for its browser alone.
Eventually, the market split in two. Though a tiny percentage of web users
sported alternative browsers including Lynx, Mosaic, Opera, and Amaya,
basically 50% of the market was using Netscape’s browser; the other 50%
was using Microsoft’s. To create “technologically advanced” sites for their
clients without alienating half the potential visitors, designers and devel-
opers felt obliged to create Netscape-specific and Microsoft-specific ver-
sions of their sites. Clients then paid more than they should have to support
the development of these incompatible site versions. Thanks in part to
protests from groups such as The Web Standards Project (www.webstan-
dards.org) and mainly to the hard work of browser company engineers, sup-
port for common standards is constantly improving—though not without
occasional backsliding.
04 0732 CH02 4/24/01 11:15 AM Page 26
26 WHY: Designing for the Medium: Web Agnosticism
Complicating the issue, many of today’s web standards were yesterday’s

proprietary innovations: things that worked only in one browser or another.
You can’t blame Wendy’s for not offering McDonald’s secret sauce, and you
can’t fault browser companies for failing to implement technology
invented by their competitors.
When Netscape unveiled <FRAMES> (the ability to place one web page
inside another), the technology was widely adopted by designers and
developers. Refer back to Figure 2.3, Assembler.org, for an example of the
way frames work. The bottom frame contains a menu; the top frame con-
tains the content. Clicking the menu changes the content by loading a new
content frame. Both frames are controlled by yet a third document, called
the <FRAMESET>, which links to the frames, establishes their size and posi-
tioning relative to one another, and determines such niceties as whether or
not the user can resize a given frame.
Eventually Netscape brought its invention to the W3C. Much later, it ended
up as part of a temporary standard: the HTML 4 Transitional Recommen-
dation. It took Microsoft a while to support frames, because Microsoft’s
browser developers had to reverse-engineer Netscape’s invention to figure
out how it worked. Ironically enough, Microsoft’s 4.0 browser eventually
supported frames better than Netscape’s.
In 1995, Netscape came up with a programming language initially called

LiveScript and eventually renamed JavaScript. Besides being easy to learn
(at least, as far as programming languages go), JavaScript made web pages
far more dynamic. And it did this without straining the computers used to
serve web pages (servers), because the technology worked in the user’s
browser instead of having to be processed by the server itself—the way Perl
scripts and other traditional programming languages had been. With less
strain on the server, more web pages could be served faster. Thus,
JavaScript was bandwidth-friendly.
JavaScript eventually became a standard, but not before putting Microsoft

at a competitive disadvantage for several years. The latest, “standard” ver-
sion of JavaScript is referred to as ECMAScript, which sounds like the noise
our Uncle Carl used to make in the morning. Don’t worry—’most everybody
still calls it JavaScript, which isn’t exactly Yeatsian poetry either, come to
04 0732 CH02 4/24/01 11:15 AM Page 27
think of it. (ECMAScript is so named because the European Computer Man-

ufacturers Association [ECMA] supervised the standardization process.)
While Netscape and Microsoft invented competitive new technologies, the
W3C worked to develop recommendations that looked beyond the
“Browser Wars.” At times, the W3C seemed to be out of touch with what
was actually taking place in the market. Back then, the browser companies
seemed to be ignoring the W3C. (The irony is that both AOL/Netscape and
Microsoft participate in the W3C and play a vital role in developing the
web standards they have sometimes gone on to ignore.) Today it appears
that the W3C is ahead of what browser companies can realistically deliver
in the next year or two. Indeed, even hardened web designers with years of
experience can feel their innards turn to jelly when reading about upcom-
ing standards proposed by the W3C. (XML Namespaces, anybody?)
The important thing is that there is now a road map for browser compa-
nies, developers, and designers. If you took your talent to the Web in the
1990s, you had no way of knowing what new technologies might come
down the pike, what new skills you would have to learn, and how quickly
what you learned (and designed) would become obsolete. Today we know
which standards have been fully or partially implemented in browsers and
which ones we can expect to work with in the next year or two. As opposed
to the past when Netscape could surprise us by inventing JavaScript and
frames or Microsoft could spring VBScript and ActiveX on us and expect us
to quickly learn and use those technologies, today we know what to antic-
ipate and what to learn to prepare for the future.
OPEN STANDARDS—THEY’RE NOT JUST FOR

GEEKS ANYMORE
We’ll bore you with the details in Part III of this book. For now, it is enough
that you understand three fundamentals of web agnosticism.
Point #1: The Web Is Platform-Agnostic

The Web owes no special fealty to any particular operating system. It is
designed to work in Windows, Mac OS, Linux, UNIX, BeOS, FreeBSD,
OS2, DOS, and any other platform that comes along. This presents web
04 0732 CH02 4/24/01 11:15 AM Page 28
28 WHY: Designing for the Medium: Open Standards
designers with special challenges in terms of gamma, screen resolution,

color palettes, and typography—all of which we’ll explore a bit later in this
chapter. This is one heck of a chapter—we hope you realize that. If you get
tired and want to take breaks, we’ll understand.
At first blush, the programmers on your team would seem to have a tougher
job than you do. How on earth are they supposed to accommodate all those
different operating systems? The answer is, they don’t have to. Browser
companies are stuck with the tough job of supporting all those platforms
(or a limited subset thereof). Web standards do the rest. JavaScript is
JavaScript whether it’s running in Linux or Mac OS. Style sheets are style
sheets whether they’re running on Windows 2000 or BeOS. The more web
standards the browsers support and the more completely they support
those standards, the fewer migraines programmers (and web users) will
have to endure.
You, on the other hand, will continually test your designs for cross-
platform feasibility. You will have to cope with the fact that your favorite
Mac system font is not available on the PC (or vice versa). That those tawny
PC colors look pale as Christina Ricci on the Mac. That the large, bold sans
serif headline that looks so dapper on systems with scalable type and built-
in anti-aliasing (such as Mac OS and Windows 98) may look hillbilly-
homely on platforms lacking those niceties (such as Linux).
What You See Is What You Get (WYSIWYG) programs, such as Macrome-
dia Dreamweaver and Adobe GoLive, attempt to give designers the sensa-
tion of retaining complete visual control over web layouts. It is an illusion.
A vast majority of professional web designers still hand-code their pages.
At the very least, they hand-tweak Dreamweaver- or GoLive-generated
code to accommodate the reality of browser and platform differences.
Browser and platform differences mean that the precise control you’ve
come to expect from publishing programs such as Quark XPress and Adobe
InDesign simply does not exist on the Web. You can bemoan this fact or
learn to create beautiful work that exploits the medium’s changeable
nature and facilitates the needs of millions—perhaps even entertaining
them in the process. Not such a bad trade-off, when you come right down
to it.
04 0732 CH02 4/24/01 11:15 AM Page 29
Point #2: The Web Is Device-Independent

Your work not only has to remain usable on a terrifying variety of computer
desktops, it also may be accessed via Palm Pilots, web phones, and other
instruments of Satan. A year ago it appeared that web designers and pro-
grammers would have to continually learn new and incompatible markup
languages to accommodate this plethora of web-enabled devices. Instead,
the W3C is guiding us toward using Extensible Hypertext Markup Language
(XHTML) and CSS to get the job done. (Don’t panic! XHTML is, more or less,
simply a newer and cleaner version of HTML.)
From www.w3.org/Mobile/Activity:
"Mobile devices are unlikely to be able to use exactly the same markup
as a normal page for a PC. Instead they will use a subset of HTML tags.
The expectation is that different devices will make use of different mod-
ules of XHTML; similarly they will support different modules of style
sheets. For example, one mobile device might use the basic XHTML text
module and the style sheet voice module. Another device with a large
screen might also allow the XHTML tables module."
The W3C website is visually lackluster, unmanageably immense, and writ-

ten in language only a Stanford professor could love. Nevertheless, the
W3C is frequently the voice of sanity in the chaos and frenzy of an ever-
changing, commerce-driven Web. Learn to overlook the site’s lack of visual
panache, and the W3C will be your best friend as the Web and your career
move forward. Which brings us to Point #3.
Point #3: The Web Is Held Together by

Standards
To design websites, you will have to learn technologies such as HTML,
JavaScript, and CSS, which really isn’t that hard. As you grow more adept,
you will become aware of wonderful features offered in only one browser
or another. We advise you to avoid these nonstandard technologies and
stick, as much as possible, to what is supported in all browsers.
You might find yourself working for companies or clients who demand spe-
cial features that only work in one browser. Just say no. On an intranet site
(see Chapter 5), it might be feasible to design a site that only works in IE5,
04 0732 CH02 4/24/01 11:15 AM Page 30
30 WHY: Designing for the Medium: Open Standards
Netscape 4, or what have you, because those who commission the site con-
trol the browsers used to access it. But we’ve heard of plenty of companies
that decided to go public with part of their intranet site—only to discover
that its nonstandard features locked out millions of web users. We also
know an agency that designed an intranet site to take advantage of
Netscape 4’s proprietary DHTML Layers technology. When Netscape aban-
doned this technology in favor of web standards, the company’s IT depart-
ment was unable to upgrade its users to the latest version of Netscape’s
browser, which would have made the site nonfunctional. Who took the
blame for this fiasco: the client who had insisted on using proprietary, non-
standard technology or the web agency that had argued against it? If
you’ve had any real experience as a designer, you’ll understand that the
question is rhetorical.
You can often get away with taking the moral high ground simply by
explaining to your clients that delivering what they request will cost them
25% or more of their potential audience. The disabled are almost always
among the first to be locked out of a site that relies on proprietary tech-
nology. Excluding millions of people from a public site is not exactly a bril-
liant business decision, and ethically speaking, it stinks. Excluding the
disabled is also illegal in many instances, at least in the United States.
Court cases have been fought over it, and the client usually loses. The Aus-
tralian Olympics website was one legal casualty; the cost to the site’s own-
ers would have wiped out poverty in three small South American nations.
If legal and ethical arguments don’t work with your clients, show them the
money.
Technologies such as HTML, JavaScript, and CSS are the building blocks of
web design. In theory, all browsers fully support these standards, deliver-
ing on the promise of browser and platform-agnosticism and offering us a
Web where we can “write once, publish everywhere.” Theory and reality
often diverge. In fact, the divergence between them is more or less the story
of the Web. The good news is that built-in browser incompatibilities are
gradually going the way of the Dodo bird as more standards-compliant
browsers become available.
04 0732 CH02 4/24/01 11:15 AM Page 31
THE 18-MONTH PREGNANCY

In early 2000, Microsoft released IE5 Macintosh edition, a browser that
delivers top-notch support for HTML 4, CSS, and JavaScript, three
immensely important web standards. Soon afterward, Opera Software
released its 4.0 browser, whose principal purpose is to deliver superior sup-
port for web standards. And a month before Christmas 2000, Netscape
delivered Navigator 6, the most standards-compliant browser yet.
To read the preceding three sentences not only induces coma, it also sug-
gests that designers are now free to use nothing but W3C standards in the
sites they and their colleagues create.
Alas, this is not the case. IE5 for Windows currently offers excellent but
incomplete support for standards. IE4, currently the most-used browser on
the Web, has good but still less complete support for standards, and
Netscape 4, still used by millions, offers even less. Sure, users can upgrade,
and eventually they will—but at their own pace.
We call this upgrade period the 18-month pregnancy, based on the time it
usually takes before web users feel compelled to switch to an updated
browser. Web designers and enthusiasts download new browsers immedi-
ately—not so your Uncle Nigel. While you beta-test next year’s browser,
your client sticks with AOL 3. Clients and other normal human beings tend
to use the browser that came preinstalled on their computers. They upgrade
when they buy a new PC. Computer manufacturers tend to install 3.0
browsers (considered stable) when 4.0 models are newly available; they
offer 4.0 browsers when 5.0 models first come out; and so on. IT depart-
ments are equally conservative, tending to view new browsers the way cats
regard changes to their litter. Those who use the Web primarily to shop,
send email, or view pornography may not be aware for months that a new
browser is available, and when they do find out they often don’t care.
The browser upgrade path is slow, thus the transition to a Web built purely
with standards could take 18 months or longer. Some say we will not see
a fully standards-compliant Web before 2003. For the near future, you will
likely find it necessary to employ nonstandard workarounds to address spe-
cific deficiencies in these older browsers. We’ll explain these workarounds
in the relevant chapters on HTML, CSS, and JavaScript.
04 0732 CH02 4/24/01 11:15 AM Page 32
32 WHY: Designing for the Medium: Chocolatey Web Goodness
Five years ago, the entire Web was a hack, held together with carpet tacks
and lasagna. We are better off now than we were then. And soon the night-
mare of browser incompatibilities will be a story we tell to bore our grand-
children.
CHOCOLATEY WEB GOODNESS

Having accepted that the Web varies from user to user and browser to
browser, and that this will be true even when common standards enjoy uni-
versal support, let’s move on to consider the medium’s many unique
strengths. If you already consider the Web the greatest thing since gender
differentiation, feel free to skip ahead.
’Tis a Gift to Be Simple

Developing effective web architecture takes great skill. Setting type,
designing images and elements, and laying out pages requires consistent
vision and intelligence. Programming sites that will serve sophisticated and
novice users alike is an art of the highest caliber. But anyone can make a
website. A child of six can learn HTML and begin self-publishing in a mat-
ter of days. No other medium is as easy to learn and produce.
Millions of personal sites prove this point. Many are of interest mainly to
their creator’s immediate family and friends—and that’s okay. But a sur-
prising number offer valuable content and/or sophisticated design. You can
view the vast outpouring of personal pages as proof that HTML is easy to
learn. You also might see in it the unshakable human urge to reach out and
connect with others. You can even view it as an extended experiment in
democracy.
Democracy, What a Concept

Every medium in human history has presented a barrier to access. Writers
have had to convince publishers that their books were worth distributing
(or else build their own printing press, like poet William Blake). Screen-
writers must convince studios to invest millions in their visions (and the
writer is usually barred from the set once the script has been sold). Movie
directors must argue with producers and bankers. Painters need galleries;
musicians need concert halls and record deals.
04 0732 CH02 4/24/01 11:15 AM Page 33
But the Web presents few such barriers. Buy a computer and a modem, find
a hosting provider, learn HTML and some UNIX filing conventions, and
voila—you are a worldwide publisher! If you can’t afford a computer and
modem, most public libraries, universities, and schools offer free Internet
access. If hosting fees are beyond your means, companies such as Geoci-
ties (www.geocities.com) provide free hosting in exchange for the privilege
of running ad banners on your site. The Web places the virtual means of
production in the hands of virtually every worker. What would Karl Marx
think?
Figure 2.5
The Stinky Meat Project.
On the Web, anyone can
publish anything they like.
Baby, that’s democracy!
(www.thespark.com/
health/stinkymeat/)
Speaking of low access barriers, remember the days when you had to
expensively laminate print proofs of your best work, slip them into a costly
portfolio, and toss them out every six months as your new work made the
old stuff obsolete? Well, forget all that. With a free or inexpensive Inter-
net account, you can mount a web portfolio that’s viewable anywhere in
the world. Nothing to replace; nothing to bang into the knees of a Nean-
derthal seated across from you on the subway; nothing for your boss to see
you lugging around when you look for a new job on your lunch hour.
04 0732 CH02 4/24/01 11:15 AM Page 34
34 WHY: Designing for the Medium: Instant Karma
INSTANT KARMA
If the invention of the printing press brought humanity out of the Dark
Ages, the building of the Internet and the growth of the Web have ushered
in a new information age. It’s an era where every voice can be heard and
where truth can win out over lies—even when the liars have million dollar
budgets. Say Detroit spews out a bad car (it happens) and decides to dump
millions on advertising in the hope of selling it anyway. Message boards on
the Web will quickly spread the word that the lemon gets five miles per
gallon and spends more time in the shop than on the road. Angry owners
may even start a protest site, garnering coverage in the traditional
news media. The Web has changed the rules of the market. (See
www.cluetrain.org for more on this.)
It also has changed publishing. Some of the Web’s best-loved authors have
never written a traditional book. Others have gotten traditional book deals
based on the popularity of their online publications.
The Web has launched careers, CDs, and movies and brought together the
globally scattered members of countless unnamed tribes. You might be the
only Sufi in Piggott, Arkansas, but you can find thousands of fellow believ-
ers online. If the other kids attending Fredericksburg High don’t share your
passion for the music of Bernard Herrmann, you’ll find folks more in tune
with your interests online.
Social commentators sometimes worry that the Web is making us more iso-
lated. In the picture these pundits paint, tortured introverts peck out des-
perate messages in dark, lonely chat rooms. We take a different view. In
ordinary life, extraordinary people often feel terribly isolated because no
one around them can understand them other than superficially. The Net
and the Web offer real hope and true companionship for those willing to
express themselves and seek out like-minded souls. This, we think, is a good
thing.
04 0732 CH02 4/24/01 11:15 AM Page 35
THE WHOLE WORLD IN YOUR HANDS

They don’t call it the World Wide Web for nothing. As individuals, we can
not only email pen pals in Istanbul and Amsterdam, we can find out what
people in those countries think by reading their personal sites or talking
with them in online communities.
People living in nondemocratic nations can publish their protests anony-

mously without fear of government retaliation. In lands where all views are
tolerated, everyone from amateur gemologists to alien conspiracy freaks
can broadcast their theories to a global audience.
Free online services, such as Alta Vista’s Babelfish (babelfish.altavista.com)

translate text on the Web into a variety of languages. These translations
may be awkward and even hilarious—after all, translation is an art best
practiced by human beings. But the gist of the text survives the transla-
tion. If you publish the story of your child’s first steps on your personal site,
your tale may be accessible to families in Indonesia and Zimbabwe.
The Web not only reaches the world, it changes it. As a web designer, you
will be an agent of change, which is a lot easier and much less dangerous
than becoming an agent of the FBI. You’ll also sleep better, and you won’t
have to wear a tie.
JUST DO IT: THE WEB AS HUMAN ACTIVITY

Unlike any other mass medium, the Web encourages human activity
instead of passive consumption. This can have a transformative effect, as
consumers become active participants, reinvent themselves as content
producers, and launch political parties or small businesses without begging
for third-party capital. Armed with nothing more than the Web, individu-
als or small groups can affect the way the world does business, call global
attention to a regional injustice, or bring hope to a cancer patient (http://
vanderwoning.com/living/blog.html).
04 0732 CH02 4/24/01 11:15 AM Page 36
36 WHY: Designing for the Medium: The Viewer Rules
Visit a web community, and you’ll see people who used to channel-surf
devoting their leisure hours to arguments, flirtations, and other classic
forms of human interactivity. These communities can spill over from the
virtual realm to the real world. The members of Redcricket, for example,
visit each other’s cities (www.redcricket.com). The readers and writers of
Fray (www.fray.org) hold live personal storytelling events each year. The
members of Dreamless (www.dreamless.org) participate in collaborative
design projects (www.kubrick.org) and hold noncommercial “underground”
design festivals in cities such as London and New York.
THE VIEWER RULES

On the Web, the viewer is in control. She can alter the size of your typog-
raphy. She can turn off images. She can turn off JavaScript. She can force
all pages to display her choice of fonts and background colors. In advanced
browsers such as Netscape 6 and IE5/Mac, she can even use her own style
sheet to disable or interact with the one you’ve designed. For designers, this
can be either a nightmare or a new way of thinking about design. The open-
minded may wish to read “A Dao of Web Design” for a positive approach
to this aspect of the medium (www.alistapart.com/stories/dao/).
Designers can thwart the user’s power if they insist—with mixed results. For
instance, to force the viewer to see what you want her to see, you can
deliver body text in an image instead of typing it in HTML. This is a classic
mistake of the novice web designer. Why is it so wrong? Let us count the
ways:
1. If the viewer has turned off images in her browser, she cannot read
what you (or your client) have to say.
2. She cannot copy and paste your text into an email message she’s
sending to her family.
3. Search engines will not see the text because it is embedded in a

graphic image, and as a result, fewer people will discover your page.
4. A near-sighted visitor might find it difficult or impossible to read

your 9pt. Futura “graphic text.”
04 0732 CH02 4/24/01 11:15 AM Page 37
5. As if those are not reasons enough to stick with HTML text, consider
the fact that each image must be downloaded, translated, and dis-
played by the browser—a process that can take more time than the
reader is willing to devote to your site.
Lest you run scared, bear in mind that most web users rarely, if ever, change
their browsers’ defaults. By default, images are turned on, JavaScript is
turned on, and style sheets are turned on, which means that your typo-
graphic choices and other design decisions come through intact (albeit fil-
tered by the visitor’s browser and platform). Nevertheless, educated users
do have the power to filter your work through their preferences, so it is
important to think of web design as a partnership with the people who read
and view your sites and to accept the fact that your layouts might be trans-
formed by visitors with special needs or quirky preferences.
Figure 2.6
An embedded Quick-
Time video at The Ad
Store’s website. QuickTime
streams the video,
enabling it to begin play-
ing before the file has
fully downloaded. In this
way, the needs of the
low-bandwidth user are
accommodated without
impacting file quality
(www.the-adstore.com).
MULTIMEDIA: ALL TALKING! ALL DANCING!

The Web not only presents text and images, it also can present music,
movies, and unique forms of interactive animation such as interactive vec-
tor animations created in Macromedia Flash (www.flash.com) and videos
delivered in the QuickTime, Real, or Windows Media Player formats. Design
benefits include the power to absolutely mesmerize viewers. Design chal-
lenges include creating the work itself; optimizing the work so that it
04 0732 CH02 4/24/01 11:15 AM Page 38
38 WHY: Designing for the Medium: Multimedia
streams quickly to the viewer’s browser instead of taking half an hour to

download; and developing alternative content for those who cannot view
the multimedia file. Additional challenges include avoiding cliches and
knowing when multimedia is inappropriate.
On the Web, multimedia is most often delivered through free players (such
as RealPlayer) and free browser “plug-ins.” These are much like the third-
party plug-ins that add new capabilities to programs such as Adobe Pho-
toshop, except that they’re free. Browser plug-ins are downloadable
mini-applications that handle specific types of multimedia (MIME) con-
tent. For instance, the QuickTime plug-in (www.apple.com/QuickTime/)
allows Mac and Windows users to view digitized videos. It also plays MP3
audio files, Windows WAV audio files, Windows AVI movies, PNG images,
BMP images, and other common media formats.
When the visitor encounters a web page with an embedded QuickTime

movie (www.apple.com/trailers/), she can watch the movie with a click of
the mouse. If she hasn’t installed the plug-in yet, she can download it and
then watch the movie. The QuickTime plug-in comes standard in both
Netscape and Explorer’s browsers, so the issue is moot for most web users,
who usually use one or both of these browsers. Flash and RealPlayer also
come standard with Netscape’s browsers.
The Server Knows

The quality of the movie may vary depending on the visitor’s access speed.
With QuickTime 4 and higher, for instance, the faster the connection, the
larger the movie and the higher its quality. This is accomplished through
an ingenious scheme whereby QuickTime content is exported (saved) at a
variety of quality levels and stored as a series of related files on the web
server. When the visitor’s browser requests the file, the server checks to
determine the visitor’s connection speed and responds with the appropri-
ately optimized file.
How does the server “know” the user’s connection speed? The plug-in
“tells” the server. QuickTime includes a control panel, which asks the user
to select her connection speed. This information is then conveyed to the
plug-in. Ingenious.
04 0732 CH02 4/24/01 11:15 AM Page 39
The server actually knows quite a lot about each site visitor’s setup. For
instance, it knows what kind of browser is requesting each file on a given
web page, which version of that browser (5.0, 5.5, and so on) is being used,
and which operating system runs on the visitor’s desktop. The ability to
access this information can be quite useful when you’re coping with
browser and platform incompatibilities—as we’ll discuss in Chapter 11, “The
Joy of JavaScript.”
Because the server finds out and records this information every time a web
page file is requested, you also can find it out for yourself by checking your
site’s referrer logs. What are those? Glad you asked.
Referrer logs are a standard means of letting the site’s builders or owners
know how many people are visiting, what browser and platform they’re
using, and which third-party sites “referred” these visitors via links. They
also track the national origin of each unique visitor, tell you which pages
are the most visited, and much more. Referrer logs are cool.
You won’t find your visitors’ names, addresses, and telephone numbers, of
course. That information is private, not because all site owners are decent
human beings but because such information is unknowable unless the vis-
itor has voluntarily supplied it.
On certain news sites (www.nytimes.com) and some database-driven sites,

the visitor must enter this private data before accessing content. The data
is then stored on a cookie on the visitor’s hard drive, allowing the user to
return to the site without having to undergo the tedious log-in process
each time. Advertisers and site owners foam at the mouth over the possi-
bility of procuring information like this. We’ve even had a client ask if there
was any way to find out each user’s business phone number “without
telling them.” (Answer: No, and if there were, we wouldn’t tell you.)
Web users’ privacy concerns make them unlikely to provide personal data
without sufficient motivation. Reading The New York Times free of charge
may constitute sufficient motivation. Finding out more about Widgets.com
probably does not.
04 0732 CH02 4/24/01 11:15 AM Page 40
40 WHY: Designing for the Medium: Multimedia
Many sites over the years have unwittingly erected barriers by forcing users
to enter personal data without first giving a clear picture of what the user
would gain by doing so. Many of these were sites flung together like so
much moldy cheese by traditional media moguls. When users failed to reg-
ister, the moguls would claim that “Web content doesn’t work” (if the ill-
conceived site was their own) or trumpet the failure far and wide (if the
site belonged to a competitor). Some of these sites offered decent content,
but few folks were willing to cross the privacy barrier to find out about it.
Though web users are understandably reluctant to reveal their salaries and
sexual preferences merely to view content, the server’s tracking of less sen-
sitive information can still be incredibly useful to the design and develop-
ment team. For instance, if you discover that a great many visitors are
coming from Sweden, you might commission a Swedish translation of the
site—thereby enticing still more Swedes to visit. If you learn that 90%
of your audience is using a 5.0 browser or better, you can incorporate
standards-based dynamic technologies with less fear of alienating your
core user base. The combination of server user awareness and sophisticated
plug-ins such as QuickTime allows you to craft the optimum experience for
each visitor.
The server can always tell each user’s connection speed, operating system,
and browser. This allows sophisticated plug-ins like QuickTime to deliver
the optimum experience for each user. As we’ll see later, it also enables us
to do clever and useful things with JavaScript.
Not every player or plug-in format accommodates user connection speed

in precisely the same way that QuickTime does. A Flash movie, for instance,
does not vary depending on the user’s connection speed; it is always the
same Flash movie. Flash, however, was designed specifically to couple rich
multimedia experiences with compact file sizes. Why is this important?
04 0732 CH02 4/24/01 11:15 AM Page 41
IT’S THE BANDWIDTH, STUPID

Aside from corporate users and a few lucky folks with Digital Subscriber
Line (DSL) or cable modem access, most people view the Web through dial-
up connections, which are not exactly peppy. In every multimedia format,
digital compression is used to compensate for the narrowness of the user’s
“pipe”—the limitations of her bandwidth. As mentioned earlier, bandwidth
represents the rate at which web content may be downloaded to the end-
user’s computer. Remember David Siegel’s cry of “Clarity, Brevity, Band-
width?” Bandwidth is arguably the most important component of this
trinity. Web users will spend more time with mediocre sites that load fast
than they will waiting for beauty that takes forever to show up on
their screens. (Q. What’s the most popular button on the Web? A. The Back
button.)
Dialup modems top out at 56K. That’s 56 kilobits, or 6 kilobytes, per

second. (Actually, it’s even less than that: The FCC mandates a top speed
of 53K. Read the fine print on your modem.) Due to modem overheads
ranging from 1% to 15%, phone line noise, server traffic levels, Internet
congestion, and the alignment of the planets, modems rarely if ever actu-
ally achieve their top speed. 33.6 modems can do no better than 4.2K per
second and frequently do less. 28.8 modems typically deliver 3 to 3.5K per
second.
In ideal conditions, under a blue moon, on the Twelfth of Never, a home

user is downloading less than 6K per second. So a 600K movie will take at
least 100 seconds to download to the user’s computer. The greater a for-
mat’s compression ratio, the fewer kilobytes (or megabytes) your visitors
have to download and the sooner they can start enjoying what you have
to offer.
Flash, RealPlayer, QuickTime, and Windows Media Player all stream their
content (begin playing the file soon after downloading begins). But even
streaming formats are limited by the bandwidth constrictions of the end-
user’s modem. Streaming or not, no multimedia format can pour its data
faster than the user’s modem can drink it.
04 0732 CH02 4/24/01 11:15 AM Page 42
42 WHY: Designing for the Medium: Web Pages Have No Secrets
As you might expect, the format that compresses best uses the least band-
width and is therefore the most popular. The RealPlayer (www.real.com) is
the “best-selling” free video player on the market because it compresses
video and audio down to sizes that work well even over dialup modems
(though 56K modems are strongly recommended). QuickTime files tend to
be larger than Real files and have higher quality; again, as common sense
would lead you to expect, QuickTime is not quite as popular as RealVideo.
Windows Media Player is currently the third most popular streaming for-
mat. Though it’s native to the Windows Operating System, an oddly named
“Windows Media Player for Macintosh” is available also, and seems to work
well enough.
When appropriate, these players and plug-ins enable designers to bring

rich multimedia (and in the case of Flash, interactivity) to the Web. And of
course, when used unwisely, they make the medium a virtual hell of ugly
spinning logos, unwanted soundtracks, and other detritus that adds insult
to injury by taking forever to download.
WEB PAGES HAVE NO SECRETS

Web pages are immodest. You can see what’s under their clothes. You can’t
learn the design secrets of a print layout by looking, touching, or clicking;
but you can easily do this on the Web.
To begin with, every browser since Mosaic, released in 1993, has a menu
item called View Source. As you’d expect, this allows you to view the source
code of any web page. How the heck did the designer pull off that intricate
web layout? View the source and find out. How did they make the image
change when you dragged your mouse across it? Click View Source and
study their JavaScript code. It is, of course, possible to obfuscate JavaScript
source code, making it difficult for source snoops to understand what is
going on. It’s also possible to write extremely ugly code, but that’s usually
not intentional. For an example of the former, use View Source and com-
pare: http://dhtml-guis.com/game/poetry.opt.html versus http://
dhtml-guis.com/game/poetry.html.
04 0732 CH02 4/24/01 11:15 AM Page 43
Naturally, you need to know enough about HTML or scripting languages to

understand the code you’re looking at. Conversely, the more source code
you view, the more you’ll learn about the code that makes web pages work.
Most web designers learn their trade this way. In fact it’s fair to say that
for every HTML book sold, there are a thousand web pages whose source
code has been studied for free. Well, perhaps it’s not fair to say, but we’ve
gone ahead and said it anyway, and since we get paid by the word, we’re
adding yet another irrelevant clause to the mess.
The ability to view source code is there for a reason: to teach HTML and
other markup and scripting languages by example. Even sharp operators
who know all the angles are constantly learning new tricks and techniques
by studying their peers’ sources.
Make a mental note never to steal someone else’s source code outright. All
you want to do is learn from it. This is an ethical and professional issue, not
a legal one. Unlike text, artwork, and photography, HTML markup is not
protected by copyright, even though some web designers claim otherwise.
Unscrupulous designers do steal each other’s code, but this is a bad prac-
tice. If the moral issues do not concern you, imagine your embarrassment—
and possible business difficulties—should your client receive an angry letter
from a designer whose code you swiped. It’s not worth the risk.
In Chapter 8, “HTML, the Building Blocks of Life Itself,” we’ll teach you how
to View Source in your HTML editor of choice rather than inside the
browser. Because many designers won’t bother reading that chapter, we’ll
pad it out with poignant childhood reminiscences and jokes involving
creamed corn.
In addition to View Source, Netscape Navigator’s menu bar offers an option

to View Document (or Page) Info. Choose it, and the entire page will be
deconstructed for you in a new window, image by image. Beside each
image’s name you’ll find its complete URL (its address on the Web), its file
size, how many colors it contains, and whether or not it uses transparency.
Click the link beside each image, and the image will load in the bottom of
the window. By viewing page info, you may discover that a large image is
04 0732 CH02 4/24/01 11:15 AM Page 44
actually composed of smaller pieces stuck together with a borderless HTML

table or that what looks like one image is actually two: a transparent fore-
ground GIF image file floating atop a separate background image. Or you’ll
discover invisible (transparent) images, used to control the spacing of ele-
ments on old-fashioned web pages. (Today, designers use CSS to accom-
plish the same thing without subverting the structural purpose of HTML.
Throw out those old web design books. The tricks they teach are outdated
and considered harmful to the future of the Web.)
Microsoft’s Internet Explorer does not let you view page info the way
Netscape’s browser does. But both browsers are free, and as a designer you
will be using both anyway. In fact, you’ll regularly be checking your work
in at least two generations of Netscape and Microsoft’s browsers and then
double-checking it in WebTV, Opera, iCab, and Lynx.
In all likelihood, even when all browsers fully support common standards,
you will still have to check your work in multiple browsers to avoid browser
bugs—and of course you will have to view your work on multiple platforms.
Or at least ask people on web design mailing lists to check it for you.
The Web Is for Everyone!

The last version of HTML—HTML 4—goes out of its way to make sure that
everyone can use the Web, from Palm Pilot owners to the blind and from
English speakers to, uh, nonEnglish speakers. HTML 4 contains improved
accessibility features that enable web designers to accommodate all
potential users, thus better fulfilling the medium’s mandate. Throughout
this book we’ll be talking about ways to make your content accessible to
everyone.
Web design is different because websites must be compatible with many

browsers, operating systems, and access speeds. The following sections dis-
cuss some of the challenges that make all the difference between design-
ing and designing for the medium.
It’s Still the Bandwidth, Stupid

In the preceding section on multimedia, we defined bandwidth in terms of
bits and bytes per second. The key to bandwidth is realizing that there is
never enough of it. Design with a few small files, and you remove the band-
04 0732 CH02 4/24/01 11:15 AM Page 45
width obstacle for most of your potential audience. Design with large files,
and your audience shrinks to a chosen few who enjoy fast access at all
times. Design with many large files per page, and your audience shrinks to
you and you alone.
Bandwidth issues are complicated by the amount of traffic clogging the

network. A corporate T1 line is very fast—until 500 employees log on over
their lunch hour. Then it can be as dreary as the slowest home dialup
modem.
Similarly, 10 early adopters share a super-fast cable modem line. They brag
to their friends who quickly subscribe to the service and tell their buddies
about it. Soon 1,000 people are connected to the same cable modem line,
and it is no longer reliably fast because the available upstream bandwidth
has shrunk. The cable modem is still offering the same peppy connectivity,
but the bandwidth is now shared across multiple users.
Likewise, an Internet Service Provider (ISP) brags in its advertising that it

offers multiple, redundant T3 connectivity (very, very fast). The advertising
campaign is so successful that a million new users subscribe to the serv-
ice, and suddenly the bandwidth available to any given subscriber is low.
ISPs are like airlines. Airlines overbook flights, causing you to miss con-
nections. ISPs underestimate needed capacity, slowing down connections.
Bandwidth never exceeds the speed of the weakest link. Your corporate T1
line does you little good if the site is being served from a home machine
connected to the Internet via the owner’s Integrated Digital Services Net-
work (IDSN) line. Or the server may be fast and powerful, but if a connect-
ing router goes down in Chicago, bandwidth will slow to a trickle.
Differences in national phone service contribute to the problem. Sites

served from Japan, Australia, and France are almost always slow to reach
the U.S. no matter how powerful the server and no matter how fast the
connection on your end.
Bandwidth also may be negatively impacted if the server is overloaded due

to temporary traffic at one of the sites it serves. In 1999, when Internet
Channel (www.inch.com) in New York City hosted a live webcast by Steve
Jobs of Apple Computer, demand for Jobs’s address ran so high that all sites
on that server ran slower than normal—even though those other sites were
unaffiliated with the Apple broadcast.
04 0732 CH02 4/24/01 11:15 AM Page 46
So let us repeat: There is never enough bandwidth. Therefore, the best web
design is that which conserves bandwidth.
Good web designers are constantly performing digital sleight of hand to

conserve bandwidth. By contrast, beginning web slingers with a back-
ground in design will typically create a comp in Adobe Photoshop, cut it
apart in Adobe ImageReady, and use Macromedia Dreamweaver or Adobe
GoLive to put it together again as a working web page. The page may look
divine, but it’s almost guaranteed to hog bandwidth.
So how do we conserve bandwidth?
Swap text and code for images

For one thing, we conserve bandwidth by using HTML text instead of typo-
graphic images wherever we can. As mentioned earlier, images must be
downloaded, decoded, and expanded in the browser —and that takes time.
Text may be downloaded in a fraction of the time. HTML is text-based and
is thus a bandwidth-friendly technology. ImageReady is a great tool, but
don’t expect it to make all your decisions for you. If you use ImageReady
or Macromedia Fireworks to generate the pieces of a web page, be prepared
to replace some of those pieces with bandwidth-friendly HTML.
Trim those image files

We also conserve bandwidth by reducing the file size of our images when
exporting them (saving them in web-friendly formats) from Photoshop. All
designers know that file sizes diminish as resolution decreases. A 1200ppi
(pixels-per-inch) image takes up more megabytes than the same image at
72ppi. On the Web, all images are rendered at 72ppi, but that is only the
beginning. Later in this chapter, we’ll discuss techniques for squeezing high
quality out of small image files, and (again) replacing images with HTML
even when you use a tool like ImageReady to automate part of the process.
Do more with less

Slicing a large image into a dozen pieces may reduce the bandwidth
required by each piece, but there is a trade-off. As the server responds to
one image request after another, the cumulative bandwidth used might be
higher than needed to serve a smaller number of larger images. Each design
requires you to experiment with these trade-offs.
04 0732 CH02 4/24/01 11:15 AM Page 47
Prune redundancy
Another technique to conserve bandwidth is to remove redundancy from
HTML code. If you’re unfamiliar with HTML, you can scan Chapter 8 for a
quick overview. But even if you don’t, the following example will probably
make sense to you. If not, just nod along and come back later.
In traditional web design, we use HTML tables to position text and images
on the page. HTML tables are just like tables in a spreadsheet, except that
the borders are usually turned off (border=”0”) to hide the underlying tech-
nology from viewers. By default, elements in a table cell are left-aligned
unless the programmer has specified otherwise by typing something like
<td align=”center”> or <td align=”right”>. Therefore, in an HTML layout, it
is unnecessary to type:
<td align=”left”>
In our code, when:
<td>
Will suffice. Now, <align=”left”> does not eat much bandwidth on its own,
but multiplied thousands of times throughout a site, that kind of unneces-
sary markup adds up to a significant waste of bandwidth per visitor. If the
site wastes 10K of bandwidth on each visitor, and one million visitors
access the site each week, the waste of bandwidth is multiplied to an
astounding 10 gigabytes per week, and visitors may experience a decline
in the overall responsiveness of the web server.
Strange as it seems, we can even conserve bandwidth by minimizing white

space in our HTML documents. Users never see these documents unless
they are utilizing View Source, and technically, the amount of white space
makes no difference in the rendering of the site. For example, this HTML
snippet:
<div align=”Center”>
<form>
<input
type=”button” style=”font-size: 12px; font-family: geneva, arial, sans-serif; background-
color: #ff6600; color: #ffffff;”
value=”Previous Reports”
04 0732 CH02 4/24/01 11:15 AM Page 48
onClick=”window.location=’com0800a.html’;”
onMouseOver=”window.status=’More of same.’; return true;”
onMouseOut=”window.status=’’;return true;”>
</form>
</div>
<p>
<br>
</p>
Is functionally identical to this HTML snippet:
<div align=”Center”><form><input type=”button” style=”font-size: 12px; font-family:

geneva, arial, sans-serif; background-color: #ff6600; color: #ffffff;” value=”Previous
Reports” onClick=”window.location=’com0800a.html’;” onMouseOver=”window.status
=’More of same.’; return true;” onMouseOut=”window.status=’’;return true;”></form>
</div><p> <br></p>
Note that this technique cannot be applied to the entire web page. If you
mess with the white space and line breaks in JavaScript, you can generate
scripting errors that cause pages to fail. It is only safe to delete the extra
white space in the HTML portion of each document. HTML does not care
whether the white space is there or not. But extra white space adds to the
character count, which in turn, beefs up the document’s overall weight. An
HTML document with plenty of white space can weigh in at 11K, while an
identical document without white space may be as little as 9K. Certainly,
2K is a negligible amount of bandwidth, but multiplied by a million users
a week as per the previous example, it once again becomes significant.
Before you rush off and start deleting all the white space from your HTML
files, bear in mind that white space helps the eye make sense of the code.
Because a site that never changes is a site that soon loses its traffic, you
will frequently find yourself reopening documents you created months
before to update the content and design. Just as often, a coworker will have
to open and revise a document you created, or you’ll be editing one of
theirs. Moreover, web design is becoming more and more collaborative,
which means more and more documents change hands throughout the
process. For this reason, most web designers leave plenty of white space in
their documents—along with a trail of comments which help the designer
or her successors make sense of the markup.
04 0732 CH02 4/24/01 11:15 AM Page 49
Typical Comments in HTML

<! -- Begin the menu bar here. -->
<! -- This script is used to preload images. -->
<! -- Another pathetic hack. -->
Bandwidth is key but not at the price of sanity. Nevertheless, some web
shops routinely save bandwidth by removing the white space from their
HTML documents. To protect themselves from suicidal despair, these shops
first save a legible copy of each document and preserve it offline. When a
particular HTML document needs to be updated, the designer or producer
opens the original document, not the one from which white space has been
removed.
Because it can be problematic and because it requires keeping duplicate

files, most shops don’t bother with this level of bandwidth conservation.
Okay, we’re sorry we mentioned the whole thing.
CACHE AS CACHE CAN

One of the best ways to minimize bandwidth is to employ the caching
mechanism built into all web browsers. The caching mechanism, which
lives on the end-user’s hard drive, is like a warehouse where files that have
already been downloaded are stored in case the user needs them again. For
instance, if a visitor returns to a previously viewed web page, the images
on that page are loaded from her cache instead of having to be downloaded
from the Web a second time. Because the files are already sitting on the
hard drive, they load almost instantly.
That’s all well and good for the web user, but how does it apply to the web
designer’s job?
The answer is simple: The more we reuse graphic elements, the less strain
we put on our visitors’ bandwidth. If we reuse the same graphic menu bar
elements from page to page, these elements only have to be downloaded
once. From then on, whenever the visitor hits a new page, the familiar
menu bar graphics are reloaded from the cache on her hard drive. By con-
trast, if we change the design of the menu bar on each page, the visitor
must download new graphics with every page, thus slowing the site expe-
rience (and adding to the toll on the server).
04 0732 CH02 4/24/01 11:15 AM Page 50
50 WHY: Designing for the Medium: Cache as Cache Can
Repetitive elements help visitors make sense of the site; ever-changing

elements confuse and disorient visitors. (Ever-changing elements don’t
help reinforce branding, either.) The need to minimize bandwidth, reinforce
branding, and present the user with a comprehensible and intuitive navi-
gational system all point to the same moral here: Keep using the same stuff
over and over, relying on the user’s cache to serve as much of the site as
possible.
Figure 2.7
The title says it all:
“a5kRobustScalableInterne
tOnlineEcommerceFurnishi
ngsOutlet,” the winning
entry in the 5k Contest,
is both a spoof and a
functioning e-commerce
site, created in less than
5K of bandwidth
(www.the5k.org/). For
those brand-new to the
field, e-commerce was the
Holy Grail of web design
in 1999.
Much Ado About 5k

The need to conserve bandwidth is so essential that in 2000, Stewart But-
terfield created a “5k Contest” challenging web designers to create some
of the smallest sites in the world: complete websites that would weigh in
at under 5 kilobytes. (To put this in perspective, 5K equals about seven or
eight short paragraphs of plain text.)
To Butterfield’s astonishment, thousands of web designers responded to

the challenge. You can see the results at www.the5k.org. As you marvel at
some of these creative solutions, bear in mind that the average web page
is 32K (over 6 times as large as the 5k winners). (The average corporate web
page is often much larger than that.) The 5k Contest proves that our pages
do not have to be nearly so bloated. As a web design professional, you will
always be seeking new ways to minimize bandwidth.
04 0732 CH02 4/24/01 11:15 AM Page 51
SCREENING ROOM
Luxuriating in your monitor’s 21” screen, you design a site that looks sen-
sational. How will it look on a 14” screen? Will it even fit? That is the chal-
lenge of screen resolution.
Screens range from 14” to 21” (and higher), with 15” and 17” currently the
most popular. By the time this book is printed, 17” screens will dominate
the home market, and ladies named Mistress Beatrice will dominate every-
where else. Laptops will continue to offer 14” and 15” screens along with
the coveted 17-incher. Not only do screens vary, resolutions vary. Some
folks view the web at 640 x 480; others at 1600 x 1200 (or even higher).
This wild fluctuation in monitor size and screen resolution has a critical
effect on page layout.
Are we saying that your site must be able to fit inside a 640 x 480 envi-
ronment? No, you don’t always have that much space. Consider that
browsers do not make full use of the screen. In Windows, room is left at
the bottom for the task bar, while the top of the screen is taken up with
browser chrome (the buttons and text entry fields that allow users to nav-
igate the Web). In Mac OS, the right-hand side of the screen is reserved for
that little trail of icons representing the user’s hard drive, saved files, and
other work-related shortcuts, and the top of the screen is again given over
to browser chrome.
Accounting for OS interface elements and browser chrome, the usable

space may be less than 580 x 380. But if you design precisely to fit that
small space as if it were a fixed newspaper ad size, your site may look for-
lorn or even ludicrous on a larger monitor running at 1600 x 1200. What’s
a mother to do?
Liquid Design
The solution is to embrace the fluid nature of the medium and, whenever
possible, design in a resolution-independent manner. Glenn Davis, web
critic and former Chief Technology Officer of Projectcool.com, uses the
phrase Liquid Design to describe an approach to web design in which the
content reflows as it is “poured” into any monitor size.
04 0732 CH02 4/24/01 11:15 AM Page 52
52 WHY: Designing for the Medium: Screening Room
Narrow your browser window to 640 pixels or thereabouts, and visit

www.jazzradio.net (see Figure 2.8). Now stretch your window as wide as it
will go (Figure 2.9). Notice how the entire layout reflows to fill the screen.
See also www.alistapart.com for another example of Liquid Design.
Figure 2.8
The original site design for
jazzradio.net works well if
the visitor’s monitor is
small…
Figure 2.9
…and equally well if the
monitor is large. Liquid
Design makes users of
any size monitor feel
equally at home
(www.jazzradio.net).
04 0732 CH02 4/24/01 11:15 AM Page 53
There are limits to how wide a web layout may be stretched before it begins
to look ludicrous, but the goal is not to provide hours of “squash and
stretch” fun for web users. (They’re not going to perform this exercise any-
way.) The goal is to provide a site that seems to naturally fit each visitor’s
monitor. This makes the visitor feel right at home, thereby encouraging her
to spend more time on the site and drink milk right out of the carton when
she thinks you’re not looking.
By contrast, with a more rigid approach to web layout, your site might
appear to be “shoved into the corner” of a user’s large monitor. Or it might
be too wide for the user’s small monitor, forcing her to scroll left and right
(or more probably, encouraging her to leave and never come back).
A great majority of websites are designed at 800 x 600 fixed resolution in

the belief that most users have screens wide enough to accommodate this
width and height. True, “most” users can accommodate it, but why not
build something that fits every user like a glove?
With Liquid Design, you can do just that.
By contrast, Banana Republic (www.bananarepublic.com) (see Figure 2.10)

and Three.oh (www.threeoh.com) offer fixed web layouts using absolute
heights and widths. Banana Republic’s site does this to fit inside small
monitors. It certainly does that, but its attractiveness is marred on large
monitors—where most of the screen lies empty and yearning.
Figure 2.10
Fixed web layouts can be
attractive, but on larger
monitors the design can
suffer from that “shoved
into the corner” feeling
(www.bananarepublic.com).
Sites must be designed to
work on small monitors but
need not be designed to
look ludicrous on large
ones. Liquid Design can
solve this problem.
04 0732 CH02 4/24/01 11:15 AM Page 54
54 WHY: Designing for the Medium: Screening Room
Where bananarepublic.com chooses a fixed layout approach to accommo-

date dinky screens, Three.oh’s large, fixed layout requires the visitor to own
a monitor big enough to take in the entire design at a glance. Three.oh is
elegantly designed and serves an audience of graphic artists. Thus, the
assumption that site visitors possess a large enough monitor to see the
whole thing is reasonable enough. But by adhering to a print-like model of
site design, using absolute widths and font sizes, Three.oh rules out visitors
saddled with small monitors as well as the visually impaired. The site’s
designers no doubt feel justified in doing this because nondesigners and
visually impaired folks could not possibly be interested in what the site has
to offer. Most sites cannot make assumptions like this.
Liquid Design is accomplished through HTML tables that are built with per-
centages (rather than absolute widths), framesets that use percentages
(rather than absolute widths), or CCS. Because 4.0 browsers are still in use
at the time of this writing and will be for at least the next year, and because
CSS support is less than perfect in 4.0 browsers, most designers choose
tables or framesets to get the job done. We’ve created a simplified HTML
example to show how Liquid Design differs from print-like, fixed design.
Peek ahead to Chapter 8 if the markup confuses you.
Traditional versus Liquid Design
Here is a traditional, print-like approach to web design that uses table cells
with absolute widths. All extraneous code has been deleted from this radically
simplified example to focus on the points of difference between print-like and
Liquid Design.
<html>
<table width=”600”>
<tr>
<td width=”400”>
<p>Content goes here.</p>
</td>
<td width=”200”>
<p>Navigation goes here. This column is half as wide as the content column.</p>
</td>
</tr>
</table>
</html>
04 0732 CH02 4/24/01 11:15 AM Page 55
Next, a similar web page, but this time it’s liquid. Specifying percentages
rather than absolute widths enables the page to fit any screen while pre-
serving the relative proportions of the original layout.
<html>
<table width=”100%”>
<tr>
<td width=”66%”>
<p>Content goes here.</p>
</td>
<td width=”34%”><p>Navigation goes here. As in the previous example, this column is
half as wide as the content column. However, this table will stretch or squash to fit any
monitor comfortably.</p>
</td>
</tr>
</table>
</html>
The liquid approach handles our horizontal problem, but what about the
vertical? Simple: Remember that the first 380 pixels of vertical space is the
only area that all your visitors are certain to see without scrolling. Make
sure that your navigational menu (if any), logo (if any), headlines (if any),
and other important content fits comfortably within that vertical area. Less
important information can fall below the fold, and no harm done. Your
client’s advertisers will be clamoring for placement at the top of the screen
for this very reason. Alas, if they get their wish, those with small monitors
will see browser chrome, ad banners, and task bars to the exclusion of
almost everything else. No wonder some people hate the Web the first time
they see it.
COLOR MY WEB
As with the wide variety in screen resolutions, computers are far from uni-
form in their ability to display color. Designers work with machines that
support millions of colors (24 or 32 bits). But many computer users are lim-
ited to thousands of colors (16 bits), and a significant minority is stuck with
256 colors (8 bits) or less.
04 0732 CH02 4/24/01 11:15 AM Page 56
56 WHY: Designing for the Medium: Color My Web
Monitors that are limited to 256 colors face an additional problem in that
up to 40 of these colors are “used up” in advance by the operating system
itself. For instance, Windows reserves 40 Windows system colors for its
own display purposes in lower-end color environments. That leaves exactly
216 colors at your disposal.
In 1994, the makers of Netscape Navigator mathematically subdivided the

color spectrum into 216 web-safe colors, which are equidistant from each
other along the color wheel. You will hear this mathematical arrangement
of web-safe colors variously referred to as the Netscape Color Cube, the
web-safe palette, and variations thereof, many of them unprintable in a
family publication.
The Color Cube is the bane of many web designers’ existence, but it need
not be. Paper stocks have limitations; so do type families, and so does the
Web. This is one of those limitations you can master upon accepting it as
part of the discipline the medium imposes.
Know the Code
Photoshop 5 (and higher) includes a web-safe color palette, and the included
VisiBone color palette is even more useful because it arranges the colors in
ways with which designers can understand and work. But how can you tell in
code alone if your colors are web-safe? Easy. Know the code. In HTML, all col-
ors are indicated in three pairs (six digits) of hexadecimal code.
This, for instance, is red: #ff0000.
And this is a darker red: #cc0000.
What are these little characters? They are hexadecimal code for the Red,
Green, and Blue channels of an RGB monitor. The first two digits indicate the
amount of light pouring from the monitor’s Red channel; the second pair tells
how much Green appears; and the third tells how much Blue.
With #ff0000, the Red channel is going full blast (#ff is the highest possible
two-digit value in hexadecimal), and the other two channels are “turned off”
(#00). (Most of the time, you will be working with subtler color values.)
Web-safe colors are composed only of the following hexadecimal pairs:
00 33 66
99 cc ff
Thus, #3399ff is a web-safe color, while #07ba42 is not.
04 0732 CH02 4/24/01 11:15 AM Page 57
Only the 216 web-safe colors (colors that can be described with the hexa-
decimal pairs indicated in the previous sidebar) are guaranteed to display
correctly in both Windows and Mac OS in the 8-bit environment. Any other
color will dither (be broken into dots) on a 256-color monitor and will shift
(change to an unintended and subtly mismatched color) on a system with
thousands of colors.
Thousands Weep
As of this writing, 56% of computer owners now have 16-bit color (thou-
sands of colors), and this probably makes them happy because it makes the
daily bikini models’ flesh tones look more realistic. But for web designers,
16-bit color is a nightmare.
Sure, the dithering in 8-bit (256-color) systems is downright ugly and can
make a web page unreadable, but you can avoid it by sticking to the web-
safe Color Cube, which thus ends the problem. By contrast, the unavoid-
able color shifting that occurs on 16-bit systems springs from the dripping
maws of Hell.
Figure 2.11
For reasons only a soft-
ware company could
explain, browsers and
image editors round off
16-bit color calculations
differently. As a result,
for users of 16-bit color,
image backgrounds and
HTML (or CSS), back-
grounds will never match
(www.alistapart.com/
stories/beyond/).
Say your web page has a web-safe, light brown background color. Say your
client’s product shot is supposed to sit on the page. Say the background
color in that product shot is subtly “off” from the background of your web
page. Say you’re in big trouble, cowboy.
04 0732 CH02 4/24/01 11:15 AM Page 58
Due to differences between the way browsers calculate 16-bit color and
the way image editors like Adobe Photoshop do it, in the 16-bit color space,
browsers are off differently from the way GIF images are off. In other
words, the background color of the image absolutely, mathematically can-
not match the background color of the web page. All the web designer’s
careful illusions are revealed. There is nothing you can do about this except
wait for 24-bit color to become cheaper so that more consumers will
adopt it.
Some web designers work around this problem by using transparent back-
grounds. This is fine as long as the image does not serve also as a link. (Most
images these days do.) Why are links problematic? Today most web pages
use the CSS hover property to make links light up (meaning change colors)
when the visitor drags her mouse over them. As you’ll see in Chapter 3, this
kind of visual interactivity is helpful because it lets the user know that this
particular set of words can take her somewhere else with a click of her
mouse. When images serve as links and when links use the CSS hover prop-
erty, the background color of a transparent image will change in response
to the actions of the visitor’s mouse. Freddie Kreuger has nothing on this
unintentional visual effect. Web designers who wish to avoid this horror
will either create incredibly complex style sheets or simply use solid, web-
safe background colors in their images. And of course, these solid colors
will be subtly mismatched on the screens of all 16-bit users. Welcome to
the Web. Meantime, at least you can protect your 8-bit, 24-bit, and 32-bit
using friends by sticking to the web-safe color palette as often as possible,
particularly for large color fields, typography, and background colors.
At this point many designers scream: “These colors are ugly! This is not
what I want.” You will find, after you work with these colors, that it is pos-
sible to create pleasing combinations with them, and you will develop your
own techniques for doing so. We promise.
When saving images, you do not need to worry about intermediate colors.
If your type is web-safe orange, and your background is web-safe blue, the
edges of the type will be filled in with intermediately shaded pixels that are
probably not web-safe. They do not have to be. As long as the large areas
of color are web-safe, a little dithering around the edges of type and
images goes unnoticed by most users.
04 0732 CH02 4/24/01 11:15 AM Page 59
While GIFs are an appropriate format for logos, typography, and illustra-
tions, the JPEG format is usually preferred for photography. It is impossi-
ble to shift colors to the web-safe palette in a JPEG. Again, this limitation
of the medium is accepted or ignored by most users. But GIF images should
generally be shifted as closely as possible toward the Color Cube. In the
next sections, we will talk about ways of doing that.
Gamma Gamma Hey!

Gamma is a measurement of light, and different platforms come with dif-
ferent standard gamma settings. The Macintosh has a System Gamma set-
ting of 1.8. Put simply, it looks bright and has a wide range of light-to-dark
variance unless Mac users adjust their display to some other setting. Sili-
con Graphics Machines (SGI) have a System Gamma setting of 2.4. Their
default output is darker than that of Mac OS.
The Windows, Linux, and Sun operating systems run on PCs. PCs and their
components are built by a wide variety of manufacturers. While this keeps
end-user costs down, it also means that PCs have no standard hardware
gamma correction. Typically, their System Gamma is estimated at 2.4—
darker than Macintosh. In practice, PC gamma can be all over the place,
but it is always darker than that of Mac OS.
What does this mean to web designers? It means that if you do not com-
pensate for this cross-platform gamma variance, the subtle “study in earth
tones” that looks so moody and mysterious on your Mac will probably look
like a “study in mud” on most PCs. Because PCs are used by at least 90%
of your audience, a study in mud is not what you want.
In the late 1990s, Microsoft and Hewlett-Packard (www.w3.org/

Graphics/Color/sRGB.html) came up with a gamma standard called
standard RGB (sRGB) that gives Windows machines a common gamma set-
ting of around 2.2—at least in theory. Of course, it doesn’t work if users
don’t select it. And if they haven’t calibrated their monitors, it still won’t
really work. But at least it gives us something to aim for. Windows-based
web designers should calibrate their monitors, set their machines to sRGB,
and find something else to worry about.
04 0732 CH02 4/24/01 11:15 AM Page 60
There are three ways for Mac-based web designers to compensate for
gamma issues.
The simplest is to download and install GammaToggle FKEY (www.acts.org/

roland/thanks/), a $5.00 shareware control panel created by Roland
Gustafsson in the mid-1990s. After it's downloaded and installed in the
System folder, this simple control panel allows you to toggle between your
Mac gamma setting and a representative PC gamma setting at the touch
of a command key. The software works flawlessly, the $5.00 shareware fee
is optional (but how could you not pay the man?), and this tool has proved
sufficient for hundreds of thousands of web designers since the earliest
days of professional design on the Web. Another advantage to Gamma-
Toggle FKEY is that it is software-independent. In other words, you can tog-
gle from Mac to PC gamma whether you’re working in Photoshop, using a
browser, or simply have the kooky urge to push a Command key in the mid-
dle of a slow day.
The second solution is to download the Furbo Filters Webmaster pack

(www.furbo-filters.com), created by Iconfactory’s brilliant Craig Hocken-
berry with kibitzing from your humble author. Unveiled in 1997, the Furbo
Webmaster pack is a constellation of Photoshop plug-ins that (among
other things) allows web designers to switch between Mac gamma and
three kinds of PC gamma. The software also lets you preview the effects of
various types of GIF and JPEG compression, and an included Web Scrubber
(based on the pioneering efforts of user interface guru Todd Fahrner) lets
you selectively shift your images toward the Color Cube. The shareware
costs $40, and may be downloaded and used indefinitely for free. A nag
screen helps your conscience decide when it’s time to pay for the software.
In 1998, Adobe got wise to this whole cross-platform gamma issue (and
related web design issues) and came out with ImageReady, a Photoshop-
like application for creating and exporting web graphics. Like Furbo Filters,
ImageReady lets you preview the effects of gamma differences and com-
pression settings on your images, and it also lets you shift your colors closer
to or further from the Color Cube.
04 0732 CH02 4/24/01 11:15 AM Page 61
In late 1998, with the release of Photoshop 5, Adobe made it possible to

compensate for gamma differences between platforms using Photoshop
alone. This is largely because Adobe supports the sRGB standard in Photo-
shop (even on Macs), and Apple supports it through the system’s included
ColorSync control panel.
Mac users, here’s how to put sRGB to work:
1. Open the RGB Settings preference in Photoshop 5 or higher and

select sRGB as your working environment.
2. Photoshop will prompt you to set up your System Gamma if you have
not done so already. In Mac OS 8 and higher, you can set your Mac’s
System Gamma to sRGB using either the Mac’s built-in ColorSync
control panel or the Adobe Gamma control panel that comes with
Photoshop.
3. Set your Mac to sRGB, and you will always be inside the Windows
gamma space. If you prefer, leave it at typical Mac gamma (or some
custom setting), and Photoshop will magically shift your images
from the Mac to the Windows color space.
Choose Your Gamma
If you continue to design for print as well as the Web, stick with Apple’s default
settings and let Photoshop toggle you back and forth between Mac and sRGB
gamma settings. If you’re biting the bullet and plunging into full-time web
design, by all means set your Mac to sRGB and be done with it. After you get
used to working inside a slightly darker color space, it will look just fine to you,
and you’ll never have to worry about gamma compatibility again.
ImageReady 2.0 is included in Photoshop 5 and higher. Photoshop 5.5 is

much more web-savvy than its predecessor, and Photoshop 6 is even more
so. We heartily recommend these later versions of Photoshop. If you use an
older version, by all means try GammaToggle FKEY or Furbo Filters.
04 0732 CH02 4/24/01 11:15 AM Page 62
62 WHY: Designing for the Medium: Typography
TYPOGRAPHY
Given what we’ve already discussed in terms of screen, color, and gamma
differences, it should come as no surprise that there are vast differences in
the way different platforms handle typography on the Web.
For one thing, different platforms offer different fonts. Two sans serif fonts,
Geneva and Helvetica, come standard with Mac OS. Geneva is not found
on any other platform, and while Helvetica is available in Linux, it may or
may not be present on Windows systems. (Arial is the standard sans serif
font that comes with Windows. There is also a version of Geneva that
PC users can download, and we believe that three or four of them have
done so.)
Confused, yet?
The 97% Solution

In 1997, Microsoft decided to do something about these typographic dif-
ferences and commissioned a set of cross-platform web fonts for both Mac
and Windows. These include Verdana, a lovely sans serif font designed by
Matthew Carter; Georgia, also by Carter, a broad-in-the-beam serif font
that can claim a distant kinship with Palatino; and Mac versions of the
Windows fonts Arial, Impact, Times New Roman, Courier New, and so on.
The notion of cross-platform web fonts was a great idea. Unfortunately,

not everyone bothered to download and install these fonts, so Microsoft
included them in its Internet Explorer browser. (That took care of all the
Windows users.) Microsoft then persuaded Apple to make IE the default
browser that comes with the Macintosh Operating System. (That took care
of the new Mac users and nearly took care of Netscape.)
This did nothing for Linux or UNIX users, but it did go a long way toward
solving cross-platform font problems because Windows and Mac OS
together make up about 97% of the market. (Depending on how you define
the market, anyway.)
That still left a huge problem unsolved: the difference in typographic res-
olution between Mac OS and Windows.
04 0732 CH02 4/24/01 11:15 AM Page 63
Points of Distinction
By default in Mac OS, there are 72ppi, and a pixel is the same as a point.
Thus 12pt. type is 12 pixels tall, 72pt. type is 72 pixels (or one inch) tall,
and so on. Of course, most Mac users set their screens to higher resolu-
tions, so this one-to-one equivalency between points and pixels soon
becomes meaningless. But 72ppi is the starting point for Macs.
Windows users start off with 96ppi resolution; thus, 12pt. type in Windows
is 16 pixels tall. Again, this varies according to the user’s choice of screen
resolutions, but 96ppi is the starting point.
In 4.0 (and older) browsers, what looks readable on a Mac looks big and
horsey on a Windows PC. Conversely, what looks tasteful and discrete on a
Windows box is often illegibly small on a Mac.
Figure 2.12
Font Wars: In 1997, CSS
expert Todd Fahrner stuck
this image in an obscure
corner of the Web. It
proved why using
points was a brain-dead
approach to CSS (too bad
so few people listened). He
sarcastically observed that
if things got much worse,
Macs would have to use
Windows-size typographic
defaults. Three and a half
years later, Fahrner’s sar-
donic prediction came true
(http://style.metrius.com/
font_size/points/
font_wars.GIF).
04 0732 CH02 4/24/01 11:15 AM Page 64
64 WHY: Designing for the Medium: Typography
Particularly since web designers began overcoming their fear of style

sheets, Windows-based designers who do not check their work cross-plat-
form have been giving Mac users type they could neither read nor enlarge
in the browser. On a PC, 8pt. type looks swell. On a Mac, it looks like 8 pix-
els, which is at least 1 pixel shy of legibility.
Year 2000—Browsers to the Rescue

In 2000, browser makers figured out how to compensate for this long-
standing problem. The first to do so was IE5 Macintosh Edition, released in
March 2000. IE5/Mac’s default setting is 16px type at 96ppi (Windows res-
olution). The Mac version of Netscape 6, released in November, followed
suit.
In IE5/Mac and Netscape 6, users can change their preferences and restore
the traditional “Mac” setting for text. By doing so, they risk continuing to
be frustrated by the typographic resolution differences between their plat-
form and the dominant Windows OS. But if they’re smart enough to change
their settings once, they’re smart enough to change them back again when
needed.
IE5/Mac also introduced text zooming, which enables users to enlarge (or
shrink) HTML and CSS text on the page, no matter how the designer has
formatted that text. This liberates web users from web designers’ mistakes
and makes the medium more accessible to the visually impaired. Netscape
6 offers similar functionality, though for some reason it was left out of the
Macintosh version (at least in the initial Netscape 6 release).
Of course, 4.0 browsers are still very much among us, and the 18-Month
Pregnancy period has only just begun. Consequently, cross-platform font
size issues will continue to plague the Web for some time to come. In Chap-
ter 10, “Style Sheets for Designers,” we’ll explain how to use style sheets
to compensate for all these incompatibilities.
04 0732 CH02 4/24/01 11:15 AM Page 65
TOUCH FACTOR
When designing a book, your choice of materials and textures is limited
only by the client’s budget. When designing a website, you have no tex-
tures whatsoever. There is no “touch factor” in work designed for the dig-
ital screen. But this lack of sensory input does not mean that the site must
be a cold, detached, clinical object. There are many tools to help you bring
humanity and warmth to the Web.
Appropriate Graphic Design

Interactivity can go a long way toward simulating the effect of the “touch.”
For instance, when you move your mouse over or press the buttons at
www.k10k.net, they seem to respond to your touch—like buttons in the real
world. Intuitive, user-centered navigation helps as well. If the architecture
is designed the way users think, navigating the site will be simple pleasure.
There will be more on all that in Chapter 3. Smart, appropriate copywrit-
ing, which reads the way people talk, also can go a long way toward bring-
ing warmth and humanity to the onscreen experience.
These approaches enable anyone to create a site that feels like a living
entity. Failure to use these tools results in a site that feels cold and dead—
high tech, but not high touch.
ACCESSIBILITY, THE HIDDEN SHAME

OF THE WEB
The framers of the Web intended it to be a medium of universal access—a
medium whose wealth of information would be accessible to anyone,
regardless of physical, mental, or technological disability. Anything that
stands in the way of that accessibility is contrary to the purpose of the Web.
It is also inhumane, and, as we alluded to earlier, it is now against the law:
04 0732 CH02 4/24/01 11:15 AM Page 66
66 WHY: Designing for the Medium: Accessibility, the Hidden Shame of the Web
Section 508 of the Workforce Investment Act (www.usdoj.gov/crt/

508/508law.html) requires all United States Federal Agencies with web-
sites to make them accessible to individuals with disabilities. Inaccessible
sites can be shut down by the government. In the private sector, inacces-
sible sites face lawsuits. In 1999, a group of blind citizens successfully sued
America Online because its service was not accessible to them.
How do you design for the blind? It sounds like a paradox, but on the Web
it is actually fairly easy.
The Web Content Accessibility Guidelines of the W3C (www.w3.org/TR/

WAI-WEBCONTENT/) spell out everything designers must do to make their
sites accessible to all.
Here are some of the things you can do to make your site accessible:
■ Your <IMAGE> tags should include <ALT> text for the benefit of the
visually impaired; adding <TITLE> attributes is a good idea as well.
<ALT> and <TITLE> attributes can be spoken by audio browsers used
by the blind, so they don’t have to miss out on any content. For
example, your web page on the wreck of the Titanic includes a pho-
tograph of that ill-fated ship. A bad <ALT> attribute reads “Image,
24K.” Well, what good is that to the disabled user? So your site has
an image, so what? A good <ALT> tag will read “S.S. Titanic.” The
<TITLE> attribute can provide additional description: “Photograph of
the Titanic on her maiden voyage.”
■ If you use frames, include <NOFRAMES> content in the frameset

document. <NOFRAMES> text shows up in browsers that cannot
view frames. Old browsers fall into this category, but so do text
browsers such as Lynx and special browsers for the blind. By copy-
ing your text and pasting it into the <NOFRAMES> area, you guar-
antee that anyone can access the information on your site, even if
he or she cannot view your spectacular visual design efforts.
■ Even if most users will be navigating via snazzy visual menu bars at
the top of your site, be sure to include simple HTML links somewhere
on the page so that the disabled—or folks with older, non-JavaScript-
capable browsers—can still find their way around the site.
04 0732 CH02 4/24/01 11:15 AM Page 67
For more on accessibility and the law, see Alan Herrell’s article in A List
Apart, “Accessibility: The Clock is Ticking” (www.alistapart.com/stories/
access/).
USER KNOWLEDGE
A website must be designed so novice users can find their way through
it without trouble. At the same time, a good site offers shortcuts and power
tools for more experienced users. How do you serve these two very differ-
ent audiences at the same time? We’ll discuss that in the very next
chapter.
04 0732 CH02 4/24/01 11:15 AM Page 68
05 0732 CH03 4/24/01 11:16 AM Page 69
chapter 3
Where Am I? Navigation
& Interface
“I LEFT MY BABY DAUGHTER in the car while I went to buy dope. Then I drove
away. I’d gone about five blocks when I realized my daughter wasn’t in the
car any more.”
So begins a brief personal narrative that fills most of the screen of a web
page. At the conclusion of this woeful tale, we see a link or button labeled
More Stories. We are likely to click it.
Before doing so, we notice that a small Narcotics Anonymous logo appears
in the upper left area of the screen and that four menu items appear in a
column on the right. The Face of Addiction, reads one. There Is a Solution,
reads another. Meetings, says a third, and Membership, reads the fourth.
Meetings takes us to a map of the United States. Clicking any city takes us
to a schedule of Narcotics Anonymous meetings in that city. The Narcotics
Anonymous logo, consistently placed at the upper left of every screen on
the site, takes us back to the first page, with its riveting personal narrative
and easily understood menu structure. Perhaps when we return to the
home page we are served a different personal story. This story may be a bit
longer than the first we encountered. After all, our attention is now
engaged because we have committed at least a few minutes of our time to
the site. At this point we are ready to involve ourselves with a slightly more
elaborate narrative.
05 0732 CH03 4/24/01 11:16 AM Page 70
70 WHY: Where Am I? Navigation & Interface: What Color Is Your Concept?
This is one possible interface for the home page of Narcotics Anonymous,
a 12-Step program that helps addicts recover, one day at a time. Recovery
begins by facing the problem and telling the truth about one’s life—how-
ever painful that truth may be. The honesty of these stories enables the sto-
ryteller to get well and his listeners to identify with the problem his story
demonstrates. The prototype web interface parallels this process because
the designers have done their homework and found out how the “product”
(Narcotics Anonymous) actually works.
WHAT COLOR IS YOUR CONCEPT?

Notice that we have not said a word about graphic design, typography, or
technology. We are simply examining a prototype whose purpose is to
immediately engage readers in the site’s drama and promise. The site
achieves this by plunging the reader into content (but not too much con-
tent) and by supporting that content with a quickly comprehensible menu
structure, as well as a linear method of reading on (More Stories).
This simple site architecture, with its emphasis on human interest, provides
an immediate way for addicts to identify with an anonymous speaker and
thus begin to admit that they suffer from the same problem. It helps the
loved ones of addicts to recognize their husbands and wives as addicts and
start to understand why Harry or Sally is “that way.” The site does not
preach, nor does it overwhelm visitors with too much initial detail. Its care-
ful structure engages the minds of a specific audience and allows them to
get whatever level of support they need.
Every site should be this effective, whether it offers help for personal prob-
lems or half-price airfare. Every site should immediately engage its
intended audience with compelling content that invites exploration. A web
designer’s first job is to find the heart of the matter: the concept. The sec-
ond job is to ensure that readers understand it too. That is the purpose of
architecture and navigation.
05 0732 CH03 4/24/01 11:16 AM Page 71
BUSINESS AS (CRUEL AND) USUAL

How would ineffective web designers and clients approach the Narcotics
Anonymous project? It wouldn’t be by providing immediately engaging
content, nor by offering a streamlined menu with both global and linear
functionality. They would likely present a standard menu bar with five to
ten choices, a tedious welcome message, stock photos of smiling families
implicitly representing addicts in recovery (at least, in the designer’s mind),
and overtly commercial tie-ins to an online retailer selling self-help books.
The interior of the site might offer similar content to that contained in our
imaginary prototype, but the content would be buried several layers down
in the site’s hierarchy, where only the most dedicated would stand a chance
of finding it. Instead of capturing and presenting the essence of the client’s
message, the site would merely mimic the boring “professional” surface
appearance of thousands of other sites. Instead of potentially saving lives,
the site would merely be one more roadblock in an addict’s troubled life.
How would cutting-edge web shops approach the project? Possibly by cre-
ating a 250K introductory Flash movie featuring a spinning hypodermic
needle. The needle might morph into a rotating navigational device. Or it
might fill with blood that drips to form letters spelling out some horrific
statistic on the mortality rate of drug addicts. Such a site might win awards
in a graphic design showcase, but it would not help a soul.
In all probability, the Narcotics Anonymous organization would never com-

mission a site like any of these, nor would we expect many drug addicts to
go online in search of help. We’ve chosen this example because it quickly
dramatizes the difference between effective and ineffective web design. In
the case of Narcotics Anonymous, it could mean the difference between
life and death. But this is equally true for any business or organization that
requires an online identity—except that what’s at stake is not the reader’s
life, but the survival of the business itself. Sites with strong concepts and
solid, intuitive architecture will live. Sites lacking those things will die.
Web design is communication. It says specific things to specific people. It

does this by offering meaningful content in the context of focused digital
architecture. Navigation and interface are the doors to that architecture.
05 0732 CH03 4/24/01 11:16 AM Page 72
72 WHY: Where Am I? Navigation & Interface: The Rise of the Interface Department
In a consumer society, communication is a function of time. Traditional

designers and art directors are trained in the art of instant communication.
They understand that consumers make split-second decisions based on
emotional responses to visual information. Which toothpaste gets tossed
into the shopping cart? A stripe of color may make one dentifrice appear
more clinically effective than its competitor. Which paperback is bought in
the airport bookstore? Color and typography make one book leap off the
shelves while another is ignored. Which of a thousand billboard messages
is remembered? The one with the smart line of copy and complementary
image lingers in the mind.
When traditional designers and art directors take their talent to the Web,
their consummate understanding of the power of the image would seem
to position them as the ideal architects of the sites they design. After all,
who knows better how to focus and deliver the appropriate message before
the consumer has time to click the browser’s Back button? In good shops,
skilled web designers are empowered to do what they do best, but this is
not the case in every web agency. Some shops constrict the designer’s abil-
ities by forcing her into a more limited role.
THE RISE OF THE INTERFACE DEPARTMENT

Traditional designers and art directors work in Design Departments and
Creative Departments. The existence of these departments indicates the
importance traditional media businesses place upon design—and rightly so.
In such businesses, designers play an essential role in the formation of con-
cepts and images that convey brand attributes and communicate mean-
ingful intellectual and emotional propositions.
Sadly, many otherwise savvy web agencies do not have Creative or Design
Departments at all. Nor do creative directors or lead designers show up
often enough on some of these companies’ organizational charts. What
they frequently have instead are Interface Departments, implicitly or
explicitly staffed by “interface designers.” This departmental label trivial-
izes and may even constrict the web designer’s potential usefulness as
brand steward, conceptualist, structural architect, and user advocate.
05 0732 CH03 4/24/01 11:16 AM Page 73
When a web designer is reduced to the handwork of graphic design, some-

body else determines the overall focus and architecture of the site. Never-
theless, the rise of the Interface Department is telling because it underlines
the supreme importance of interface design to web development.
Designing interfaces is only part of a web designer’s job in the same way
that working with actors is only part of a movie director’s job. A director
who can’t work with actors will make a lousy movie, and a web designer
who can’t devise the most communicative interface for each particular site
will serve up mediocrity. Websites provide content; interfaces provide con-
text. Good interfaces support the visitor’s (and client’s) goals by visually
and structurally answering two urgent questions:
1. What is this? What kind of site is this? What is its purpose? What
messages are being conveyed or services offered? For whom is this
site intended? If it’s intended for me, does it offer the product or
information I’ve been seeking, or is it all show and no substance?
2. Where am I? What kind of space is this? How does it work? Can I

find what I need? If so, can I find it quickly? If I take a wrong turn,
can I find my way back?
When a web designer fully understands the nature of the product or serv-
ice, as in the example of the Narcotics Anonymous prototype above, then
content and context, meaning and architecture, are one. Not only does the
Narcotics Anonymous prototype quickly reveal the site’s purpose by
emphasizing appropriate text, it also understands and fulfills its potential
viewer’s gut-level needs by functioning simply and transparently. A wife
who fears her husband is becoming an addict does not have time to waste.
If the site confuses her, she’s gone.
When a web designer does not fully understand the nature of the product
or service—or understands but is not empowered to act upon that under-
standing—we get sites that excite and engage no one. Or we get poten-
tially engaging sites that confuse and estrange the very people they worked
so hard to attract.
05 0732 CH03 4/24/01 11:16 AM Page 74
74 WHY: Where Am I? Navigation & Interface: Form and Function
There are too many such sites on the Web. What businesses must under-
stand is that vague, non-engaging interfaces are a death sentence because
they alienate potential readers, members, or customers rather than reas-
suring them that they’ve come to the right place. Good web design plunges
the visitor into the exact content appropriate for the most efficient (and
personal) use of the site and continues to guide him or her through each
new interaction.
Movies immediately plunge a protagonist (and the audience) into conflict

and action. Entertainment sites can work the same way.
Newspapers carry many stories but call the reader’s attention to the most
important ones. Content sites can work the same way.
Stores sell many products, but special displays on featured products arrest
shoppers’ attention as they enter. Commercial sites can work the same way.
FORM AND FUNCTION

Effective interfaces not only lead visitors to the content but also under-
score its meaning, just as chapter divisions underscore the meaning of a
book’s content. Without usable, intuitive interfaces, websites might as well
offer no content at all—because no visitor will be able to find it.
At their most basic level, web interfaces include navigational elements

such as menu bars, feedback mechanisms such as interactive forms and
buttons, and components that guide the visitor’s interaction with the site
such as magnifying glass icons and left or right arrows. Tired interfaces
offer exhausted metaphors such as the ubiquitous folder tab and the
heinous beveled push-button. Better interfaces are uniquely branded and
help reinforce the site’s thematic concerns (see Figure 3.1).
The Mary Quant site is a study in quick visitor orientation and structurally
grounded design. the dominant but fast-loading photograph telegraphs
“1960s” and “mini-skirt,” which are the essence of fashion designer Mary
Quant’s legacy. The flower motif reinforces the 1960s theme as well as
Quant’s identity. A large flower fills in the space behind appropriately min-
imal text content; this is a fashion site, not a Ph.D. dissertation. Smaller
flowers brand the five simple structural divisions: History, Makeup, Press
Office, Shops, and Homepage.
05 0732 CH03 4/24/01 11:16 AM Page 75
Figure 3.1
The Mary Quant site—the
perfect combination of
solid design and ease of
use (www.maryquant.co.uk).
The History label is faded to reinforce the visitor’s position within the site’s
hierarchy. The Previous and Next buttons are placed left and right where a
western audience would expect them and where even non-English speak-
ers (at least those who read from left to right) will likely understand what
these buttons do.
Although this is a fashion site, its structure is nearly identical to that

sketched out in our imaginary Narcotics Anonymous prototype. The Previ-
ous and Next buttons provide linear navigation. Menu icons let the visitor
jump from section to section. Engaging visual and text content match the
desires of the intended audience.
Sophisticated interfaces work on multiple levels. On a well-made catalog

site, not only will visitors find a main navigation bar, they also will be
guided by contextual, user-driven navigational elements throughout the
page. Both the photograph and the text description of a blue parka can
serve as links to more detailed photographs and information or to an order
form. The product photo caption may include a link to More Items Like This
One, initiating a new and more focused search. Navigation does not live by
menu bars alone.
05 0732 CH03 4/24/01 11:16 AM Page 76
76 WHY: Where Am I? Navigation & Interface: Form and Function
Figure 3.2
Multi-level navigation in
action: the Gap site pres-
ents visitors with an over-
all menu bar but does not
limit them to it. Clicking
the model’s photograph…
Figure 3.3
…links the visitor to a
page displaying the jacket
the model is wearing,
along with relevant text
information and the
opportunity to buy the
item (www.gap.com).
05 0732 CH03 4/24/01 11:16 AM Page 77
COPYCATS AND PSEUDO-SCIENTISTS

A site’s navigational interface is the leading edge of the visitor’s experi-
ence. It facilitates human needs or thwarts them. If it is not intuitive, it is
useless. One reason we have so many unimaginative interfaces (visual
Muzak) is because their familiarity makes them appear intuitive, and they
therefore survive the pre-launch “user testing” phase.
For several years, nearly all sites offered left-hand navigation (menu items
on the left side of the web page, content on the right). Was left-hand nav-
igation easier to use or understand than any other configuration? No. In
fact, some studies suggested that navigation worked better on the right.
Navigation cropped up on the left because it was easier for web designers
and developers to create HTML that way—and later, it was easier to con-
trol <FRAMES> that way.
Because it was easier to program, a few large sites such as CNET.com

began offering left-hand navigation. Since CNET.com was a successful site,
unimaginative web agencies copied its interface in hopes that CNET’s suc-
cess would somehow rub off on them. With so many sites engaging in this
practice, consumers got used to it. Thus, in unsophisticated user
acceptance testing, left-hand navigation was considered “intuitive”
because consumers were accustomed to seeing it—not because it had any
intuitive advantages on its own. The “folder tabs” metaphor used at Ama-
zon.com has been copied for the same reasons. Every Nike spawns a thou-
sand swooshes; every successful site with a particular stylistic flourish
leaves a hundred thousand imitators in its wake. Bad processes encourage
bad design.
There are good marketers and there are dolts in suits. Similarly, there is
good user acceptance testing and there is worthless pseudo-science that
promotes banality. Unfortunately, worthless pseudo-science is as easy to
sell to web agency CEOs as it is to clients. It’s hard to tell until you’re actu-
ally working at a web agency whether its testing practices are informative
or a shortcut to Hell. An engaged and thoughtful web designer will develop
and fight for the best navigational structure for each site, knowing that
each site is unique because its content and audience are unique.
05 0732 CH03 4/24/01 11:16 AM Page 78
78 WHY: Where Am I? Navigation & Interface: Chaos and Clarity
Figure 3.4
Ye Olde Left-Hand Nav Bar
in action, seen here on the
Winter 2000 edition of
Icon Factory, creators of
free, funky Mac desktop
icons since 1995
(www.iconfactory.com).
The left side is no better or
worse than any other
menu placement. But for
several years, nearly all
sites stuck their menus on
the left because, well,
nearly all sites stuck their
menus on the left. Most
left-hand navigation bars
are nowhere near as cute
‘n cuddly as Icon Factory’s.
CHAOS AND CLARITY

Beyond providing access to and subtly reinforcing a site’s content, the
interface also enables people to engage in interactive behaviors, such as
shopping and searching. Or it frustrates them and sends them scurrying to
a competitor’s site, as in Figure 3.5, where clutter and lack of differentia-
tion create chaos rather than a satisfying user experience. Sites of this
nature, if they do not die immediately, persist in spite and not because of
their architecture. They survive by offering something of value to those
who are willing to overlook the experience’s deficiencies. With better
architecture they would attract more customers.
05 0732 CH03 4/24/01 11:16 AM Page 79
Figure 3.5
Where do I go from here?
Most likely, my browser’s
Back button. Busy inter-
faces bore or confuse
all but the most die-
hard bargain seeker
(www.overstock.com).
We once inherited an entertainment site that worked only on one platform

and one browser (no names, please). Our client pointed out that he was
getting four million visits a month. We replied that he was cheating him-
self out of an additional million visitors. Similarly, the owners of cluttered
and confusing sites frequently mistake a profit margin for success. Better
user experiences mean bigger profits, which is the best way to sell them to
clients whose sole concern is money.
Clients are not alone in sometimes forgetting that sites are created to serve
human needs. Web designers also can lose sight of their work’s primary
objective.
05 0732 CH03 4/24/01 11:16 AM Page 80
80 WHY: Where Am I? Navigation & Interface: Chaos and Clarity
A Design Koan: Interfaces Are a Means too

Often Mistaken for an End
As web designers become expert at crafting more and more sophisticated
navigational structures, we sometimes forget that our interfaces do not
come into being for their own sake. Interfaces are built to serve the user,
not to demonstrate our cleverness and technical mastery (unless cleverness
and technical mastery are an essential part of the brand). The best design
may go unnoticed by users, but Heaven is watching and you will get your
reward.
Universal Body Copy and Other Fictions

Good copy comes from the product; good interfaces come from consider-
ing the particular audience, content, and brand attributes of each site.
When navigation anticipates the visitor’s needs and guides her through the
site, it succeeds at the baseline level. When it does this in a fresh and
brand-appropriate manner, it succeeds as effective web design.
In this sense, web design is no different from advertising, print, or product

design. At the lowest level, an advertisement’s text must be grammatical,
and its presentation must be legible. At the highest levels, design and con-
cept are indistinguishable from the product experience. (Many would say
they are the product experience.)
Impeccable graphic design does not necessarily equate to good interface

design. As suggested by the design koan above, a site that looks drop-dead
gorgeous but confuses visitors is a site that fails.
At the turn of the Millennium, several high-stakes web businesses went

under because they forgot that their interfaces were supposed to be used
by human beings. Looking at comps and demos, the board members said,
“Oooo-Ahhh!” But when attempting to navigate the completed sites, the
public went, “Huh?” The public is the final court of appeals.
05 0732 CH03 4/24/01 11:16 AM Page 81
There were other reasons a number of web businesses failed in late 1999
and early 2000. Some businesses that served no earthly purpose and
appealed to no imaginable audience managed to suck up venture capital
anyway—until the investors woke up. But many sites with legitimate busi-
ness models bit the dust when it was discovered that nobody could navi-
gate them except, perhaps, the designers.
Each site speaks to a particular demographic. A site that is “everybody’s

friend” is nobody’s best friend. Focused, usable, brand-supportive inter-
faces are as particular as the taste of a fresh-picked plum on a summer’s
day.
While great web design, like all great design, is specific in nature, web
design (like all design) has developed a series of guidelines and best prac-
tices that can aid you as you begin to shape your own sites. Some of these
practices are rooted in common sense, others in human interface guide-
lines developed during the personal computer revolution of the 1980s. We
will examine these guidelines in the following sections, bearing in mind
that they are suggestions, not rules.
Interface as Architecture
Navigation is the experiential architecture of a site. Web designers use
consistent visual cues to guide visitors through the site, as an architect
guides a building’s visitors from the lobby to the elevator bank. Subtle
visual hints cue a building’s visitors as to which areas of an office are open
to the public, and which are private. Folks can find their way to a bathroom
or a public telephone without asking for help. The goal of a navigational
interface, like the goal of real-world architecture, is to enable people to do
what they need to do.
As you develop web interfaces, ask yourself if you’re helping people find
the site’s offices, elevators, and bathrooms or leaving them to fend for
themselves. Poorly structured buildings win few tenants; poorly structured
sites win few repeat visitors.
05 0732 CH03 4/24/01 11:16 AM Page 82
82 WHY: Where Am I? Navigation & Interface: Ten (Okay, Three) Points of Light
TEN (OKAY, THREE) POINTS OF LIGHT

In her book, Web Navigation: Designing the User Experience (available at
www.oreilly.com/catalog/navigation/), Jennifer Fleming describes ten qual-
ities shared by successful navigational interfaces. Fleming’s ten points defy
quick summarization, so we’ll settle for three of them. In Fleming’s view,
good interfaces should:
1. Be easily learned
2. Remain consistent
3. Continually provide feedback
Be Easily Learned
A designer who buys Adobe Illustrator will accept the product’s learning
curve; an online shopper will not invest the same kind of energy into
figuring out how www.halfpricefurniture.com works. Overly complex inter-
faces may please the designer who came up with them, but they rarely win
favor with those trying to find their way through the site.
Why do most of us hate the remotes that come with our TVs and VCRs?
Because there are too many buttons to push, and there is rarely an intu-
itive logic to the placement and size relationships of these buttons. We are
always hunting for the button that resets the clock or programs the chan-
nels (and discovering that this function actually lies buried deep in a series
of onscreen menus). We approach even the most basic tasks with the sense
that we are somehow being forced to prove our mastery over a trouble-
some object.
Unless we wish to watch one TV channel for the rest of our lives, we have
no choice but to click our way through the madness. But web users always
have a choice—they can visit a website that is easier to use.
Remain Consistent
Each site presents the visitor with a unique interface. Compelling content
or useful services are the only reason users bother learning how your site
works. After they’ve gone to that trouble, they will not appreciate your
changing the interface, misguidedly groping after “freshness.”
05 0732 CH03 4/24/01 11:16 AM Page 83
Web users are not mind readers. After they’ve learned that flowers serve as
visual links (as in Figure 3.1), you’d be foolish to switch to a folder tab
metaphor. If there are five main menu items per page, suddenly adding a
sixth and seventh at the same hierarchical level could make naïve web
users think they’ve somehow linked to an unrelated site. Sophisticated
users will think the site is being redesigned, and they’ve somehow caught
you in mid-process.
Many times beginning web designers feel that each section of a site
requires its own distinctive signature. It usually makes more sense to pro-
vide a consistent interface, acknowledging the new section (if at all) with
a subtle color change or a simple section title.
Figure 3.6
Digital Web Magazine, a
popular online resource for
web designers, offers a
consistent interface
between sections…
Figure 3.7
…but differentiates each
section with a subtle color
change. Because you can't
tell that the color is
changing with the color
scheme of this book,
you'll have to visit the
site and see for yourself
(www.digital-web.com).gure
3.5
05 0732 CH03 4/24/01 11:16 AM Page 84
84 WHY: Where Am I? Navigation & Interface: GUI, GUI, Chewy, Chewy
Continually Provide Feedback

In Chapter 2, “Designing for the Medium,” we remarked on the “look and
feel” issue and discussed a major difference between print and the Web. On
the computer screen, there are no matte or glossy papers, no subtly tex-
tured finishes, no chance for the designer to emboss or overprint to achieve
a richer look.
But what we lack in ink and paper choices, we make up for with an almost
limitless variety of interactive options. On the Web, the images we create
can respond to the visitor’s virtual touch. This not only adds richness to our
design; it helps the visitor comprehend the interface.
In the real world, buzzers buzz and doorknobs turn. Good web design mim-
ics this kind of feedback, using techniques such as the JavaScript rollover
(image swap) to create a sense that the site is responding to the visitor’s
actions.
Such digital responsiveness is nothing new. It began with the desktop com-
puter revolution and specifically with the Apple Macintosh Graphical User
Interface (GUI).
GUI, GUI, CHEWY, CHEWY

A website’s GUI includes all its non-text visual elements. The GUI allows
users to perform actions by interacting visually with the various graphical
elements. Familiar GUI elements from the Macintosh Operating System
include file and folder icons, scroll bars, and the Apple Menu. Windows has
its own unique GUI with elements such as the Task Bar and Start Menu.
If you were still awake a few paragraphs above when we made the big stink
about consistency, hold your nose ‘cause here we go again. Logic and con-
sistency are two reasons that Windows, Mac OS, and other UI-based com-
puting systems are more popular than command-line interfaces. GUIs
succeed by being clear (users don’t wonder what a certain button does) and
remaining consistent (if the File menu is on the left, it stays on the left).
Because your visitors are using a computer to view your site, they expect
such consistency.
05 0732 CH03 4/24/01 11:16 AM Page 85
It is worth studying existing GUIs (such as Mac OS and Windows) to figure

out what their conventions are and why they work. If your GUI works in
similar ways, you are that much less likely to baffle your audience.
Figure 3.8
The interface at panic.com
not only suggests the
Macintosh GUI, it actually
emulates it. Because the
site hawks Macintosh
software, the emulation
reinforces the site’s
themes and purposes.
Mac users will think it’s
fun; Windows users will
go somewhere else—
appropriately, since there
is nothing for them here
(www.panic.com).ure 3.5
It’s the Browser, Stupid

On the Web, the browser predetermines many elements of the GUI. For
instance, in nearly all browsers, dragging a mouse cursor over a live link
causes the cursor to change from an arrow to an upraised hand. These
browser-based conventions help web users make sense of sites. Folks rely
on these elements to understand what is happening without having to
learn an entirely new set of conventions each time they load a new URL.
05 0732 CH03 4/24/01 11:16 AM Page 86
86 WHY: Where Am I? Navigation & Interface: GUI, GUI, Chewy, Chewy
Web designers can change or override these conventions—for instance, by

using Cascading Style Sheets (CSS) to place a hand cursor over plain text
rather than live links—but it is rarely desirable to do so unless your goal is
to confuse your visitors. There are sites, such as www.jodi.org and
www.superbad.com, whose purpose is just that. These fall under the head-
ing of fine art, and many web designers adore them. Even if they’re not to
your taste, you can learn a great deal about web users’ expectations by
studying the way these sites subvert them. On most sites, though, confus-
ing visitors is usually not among the client’s objectives.
Though the browser creates many GUI elements (underlined links, changes
to the cursor state), the rest is up to the designer. Indeed, in a graphical
browser, one could consider commercial sites custom GUIs whose purpose
is to enable visitors to perform tasks while subliminally absorbing the
client’s brand.
Figure 3.9
Visitors know what this
cursor change means
(www.glish.com).
05 0732 CH03 4/24/01 11:16 AM Page 87
Figure 3.10
So why confuse them
with this one? Changing
familiar GUI elements
“because you can” is a
dog’s rationale for licking
himself. In this case, it’s
a Glassdog’s rationale
(www.glassdog.com).
CLARITY BEGINS AT HOME (PAGE)

In developing GUI elements, web designers will frequently begin with the
brand: funky elements for an entertainment site pitched at 20-somethings;
somber, restrained elements for a news or medical site; and so on (more
about branding in a moment). As each site presents a visitor with new GUI
elements, those elements have the potential to brand the site while offer-
ing visitors a sense of identity and place. These elements also have the
potential to confuse the heck out of people. As with the operating systems
they mimic, GUI elements should be as clear and easy to use as possible.
Clarity and ease of use are especially crucial factors in the development of
iconic interface elements and site structure labels.
05 0732 CH03 4/24/01 11:16 AM Page 88
88 WHY: Where Am I? Navigation & Interface: Clarity Begins at Home (Page)
I Think Icon, I Think Icon

Graphical devices (icons) guide viewers through the site experience. For-
ward and reverse arrows are common ways of navigating from page to
page. Graphical buttons are often used to trigger certain actions. For
instance, a Play button may be used to trigger a recorded sound or an
embedded, streaming QuickTime movie. A pen or pencil icon may link to a
message board, or a book or newspaper icon can guide the visitor to a
downloadable, printer-friendly version of the page’s content.
Printing in the Browser Wars
Why aren’t web pages themselves printer-friendly? It is because too often

browsers are rushed into production as the latest assault in the “Browser
Wars,” instead of offering carefully considered and usable features. By the
time this book is released, the worst of the Browser Wars will be behind us.
Icons, with or without text labels, frequently serve as quick, visual cues to
the site’s offerings. They also support international visitors for whom Eng-
lish is not a first language. Sites with massive amounts of content on their
home pages, such as portals and magazine sites, can use icons to better
organize and clarify sections (see Figures 3.11 and 3.12).
Figure 3.11
The icons seen here help
draw the eye to the sec-
ondary menu, and some of
them even communicate
in ways a non-English
speaking visitor might
understand. Designing
icons that communicate is
difficult. Competing ele-
ments must fit within the
narrow width of a lowest-
common-denominator
monitor, leaving little
room in which to develop
legible imagery
(www.eloquent.com).
05 0732 CH03 4/24/01 11:16 AM Page 89
Figure 3.12
We are clearly in the land
of the recreational web-
site, as denoted by the
tagline “professional
martini consumer.” Few
sites would devote all
that screen space to a
menu structure. Indeed,
this site recently went
offline for a redesign
(www.drymartini.com).
On the Web, as in talking to a policeman, clarity is a virtue. While it is

tempting to get really creative with such elements, the most creative solu-
tions are often the clearest.
Say you are designing a site for a chain of Wild West theme hotels. In vis-
iting the hotels and studying the chain’s promotional brochures and adver-
tising, you can’t miss the fact that Western paraphernalia is used to brand
the franchise—from the bronze horse-head coat hooks in guest closets to
the cowhide couches in the lobby. Thinking like a brand steward, you decide
it might be fun to use lassos rather than arrows to indicate “previous page”
and “next page” on the site. To you, as a visual person, it is readily appar-
ent that the rope at the edge of the lasso “points” forward or backward.
Well, cowboy, test that design on some users before you fight for it. If users
are confused by your branded iconic elements—if the lassos strike users as
meaningless ornamentation rather than functional GUI elements—be pre-
pared to rustle up some traditional left and right arrows, even if it chaps
your spurs.
05 0732 CH03 4/24/01 11:16 AM Page 90
Adding “invisible” text labels to an icon via the <ALT> attribute of the
HTML image tag or the <TITLE> attribute of a linked image can help explain
the icon’s purpose to inexperienced users. In modern graphical browsers,
these <ALT> and <TITLE> attributes generate popup “tool tips” or help-bal-
loon-style blurbs, enhancing the page’s interactivity in a meaningful and
user-friendly way.
Such tags also make the content more accessible to the visually disabled,
to those using non-graphical browsers or Personal Digital Assistants (PDAs)
such as the Palm Pilot, and to folks using conventional browsers who surf
with images turned off. (As mentioned in Chapter 2, accessibility makes
good business and moral sense. Besides, it’s U.S. law.)
When invisible text labels are not enough, consider adding visible text.
Structural Labels: Folding the Director’s Chair

In the early days of the Web, designers and copywriters frequently had fun
coming up with creative labels for menu bar sections and other naviga-
tional items. For instance, the home page of a video editing company’s site
might be labeled “The Director’s Chair,” while downloadable video clips
would be found in “The Screening Room.”
Today, most web agencies find it better to err on the side of clear copy than
cute copy. After all, if the visitor does not immediately grasp what “The
Screening Room” means, she could leave the site without having discov-
ered one of its most important content areas. While alternatives to tradi-
tional labeling may be appropriate for some types of sites (gaming sites,
fun sites for kids), many corporate sites depend on such traditional labels
as Home, About, and Clients to facilitate easy user navigation. Dull as dish-
water, we know. Be creative clearly, and it need not be dull at all.
The Soul of Brevity

Back in Chapter 2 we recalled David Siegel’s three hallmarks of good web-
site design:
■ Clarity
■ Brevity
■ Bandwidth
05 0732 CH03 4/24/01 11:16 AM Page 91
Because most web users have little time and less bandwidth to waste, good
interfaces are rarely overwrought. Given the choice between a simple,
functional design and one that is ornate, most folks prefer the simple web
layout that loads quickly and is easy to understand. Web users don’t tell
you this by peering over your shoulder; they tell you this by visiting the site
or neglecting it.
Even when bandwidth is not an issue, quick, clear communication always

will be. Users lucky enough to have T3, cable modem, or DSL access may
not be slowed down by a cluttered interface, but they will be just as baf-
fled by it as dialup modem users are. Regardless of the user’s access speed,
your communication must be fast and clear, or users will retreat faster than
you can say “failed dot com.” It’s a peanut butter and jelly scenario: By
focusing on functionality, you will develop low bandwidth interfaces; by
focusing on bandwidth, you will develop interfaces that speak quickly and
clearly.
Many web designers initially feel constrained by this. Some feel they can-
not truly express their vision unless every page sports a 128K background
JPEG, an animated menu bar, and a series of spinning logos and pulsing
photographs. We’ve all had that feeling. It passes as you discover the joy
of communicating richly while using a few elements well, or it never
passes, and you locate clients with tastes as baroque as yours. When citi-
zens avoid visiting the resulting sites, your client and you can toast your
superiority to the rest of humanity and then hurry on to the next failure.
When bad web designers die and go to Hell, they will spend eternity search-
ing for the Heaven option on an endless menu bar of purgatories. (That is,
if they’re not simply stuck waiting for an infernal intro to finish down-
loading.)
Hypertext or Hapless Text

Brevity is just as important when putting text content on the Web. A book
is easy to read. Hundreds of years of book design make it so. But on a glar-
ing computer screen, at 72ppi (pixels per inch) or 96ppi, reading long pas-
sages is a chore. A reader will simply skip lengthy texts, whether they’re
providing valuable product information or explaining how to use some
advanced feature of the site.
05 0732 CH03 4/24/01 11:16 AM Page 92
By breaking text down into usable sub-units of information, a web designer

can help readers find critical information and more easily absorb content.
White space, while useful in print, becomes even more crucial on a web
page. The logical separation of chunks of information helps engage read-
ers and maintain their interest. Designers can use paragraphs, section
breaks, and links to new pages to chunk information.
The more white space, the greater the chance that readers will remain
engaged. Use CSS by itself or in combination with table-based layouts to
create pages that demand to be read.
Figure 3.13
Readable typography, an
elegantly spare layout, and
plenty of white space add
up to a site that welcomes
readers—a quality that is
depressingly rare on the
Web (www.harrumph.com).
Contrast this with
Figure 3.15.
In print, a designer might include ten sentences in a paragraph. On the

Web, with its scrolling interface, ten sentences can feel like a life sentence.
To enhance readability, web designers (or web designers in combination
with web-savvy copywriters and editors) will separate one long paragraph
into several shorter ones.
Learn when to stop one page and start another. Despite what some pun-
dits tell you, readers will scroll to read an engaging story, but they will not
scroll forever. After two or three screens, it may be time to present the
05 0732 CH03 4/24/01 11:16 AM Page 93
reader with an arrow (or other page indicator) allowing them to move on
to the next page of text. Doing so can relieve eye fatigue, enhance the
drama of the presentation (www.fray.com), or simply give your client
another page on which to sell ad banners.
Remember in Chapter 2 when we talked about the tradeoff between one

large image that takes a long time to download and many small images
that take a long time to display? (If this were a web page, we’d provide a
link here.) Well, the same kind of tradeoff goes on with text. Jam too much
of it on a single web page, and readers may be frightened away. Provide
too little, forcing the reader to click to a new screen after every paragraph
or two, and you practically guarantee that no one will read to the end of
the article or story.
Working with client-supplied text is particularly tricky. If average citizens

are bad writers, clients are bad writers with egos. Upper Middle Managers
would rather add value to cross-brand synergies while enhancing the func-
tionality of strategically targeted product from the dairy side than put milk
in their coffee. Rare is the client who writes the way people talk; rarer still
is the client who uses few words when many will suffice.
In brochures and catalogs, such copy is ineffective. On a web page, it’s

destructive on a nuclear scale. Consumers may ignore bad catalog copy if
the layout and photography are compelling enough. But a site laden with
vast blocks of ham-handed text is doomed. No visitor will stay long enough
or scroll far enough to discover the million dollar photographs or com-
pelling brand proposition buried on page three.
Laid out well (via text chunking and CSS), bad text can squeak by. Laid out
badly, it kills websites dead. We cannot overemphasize the impact (and
tragic rarity) of good writing on the Web nor the harm done by verbose and
inexpressive texts, drizzled into layouts like so much phlegm. Learn web
typography, practice text chunking, and work with good writers and edi-
tors. Do not let your clients or your project managers skimp on the writing
budget unless you find failure exciting.
05 0732 CH03 4/24/01 11:16 AM Page 94
Figure 3.14
The front page
of Sapient.com
(www.sapient.com),
a leading web agency,
shows mastery and promise.
Clean typography and high-
quality photography, bal-
anced as skillfully as in a
classic Ogilvie print ad,
direct the visitor’s attention
to the most important con-
tent. The carefully balanced
page also makes use of
Liquid Design (see Chapter
2) to accommodate
variously sized monitors.
So far, so good.…
Figure 3.15
…Alas, once past the front
page, visitors encounter too
many pages like this one,
where blocks of undifferenti-
ated text, laid out with little
care and no love, beg to be
ignored rather than read.
Since 99% of the Web con-
sists of text that is intended
to be read, the lack of atten-
tion to good textual presen-
tation is tragic—hurting not
only the site owner, but the
would-be reader. Contrast
this with Figure 3.13.
(www.sapient.com).
05 0732 CH03 4/24/01 11:16 AM Page 95
Scrolling and Clicking Along

Some “experts” will inform you that users don’t click. They also will inform
you that users don’t scroll. If users never clicked or scrolled, nobody would
actually be using the Web. Of course users click. (How else would they link
from page to page?) Of course users scroll. (How else would they, uh,
scroll?)
Nobody clicks more than they have to—hence the so-called “Three-Click
Rule,” described later. And nobody scrolls for fun and profit. Visit an ama-
teur home page and see how excessive scrolling drags its nails across the
blackboard of the user’s experience.
The previous section, “Hypertext or Hapless Text,” discussed text chunking

and offered methods to keep scrolling to a minimum, but this does not
mean that every web page should be limited to one or two paragraphs of
text. Particularly when presenting in-depth articles online, text chunking
has its limits. Users would probably rather scroll through five longish pages
of text than click through 25 short screens that present the same infor-
mation. Develop a case-by-case, site-by-site sense for these nuances, and
you will find your skills in demand.
Every newspaper is designed so that the most important headlines, photo-

graphs, and stories appear “above the fold” (where the paper naturally folds
in half). As shown in Chapter 2, vital information is best served in this
small space above the fold. When links to the site’s most important con-
tent appear within the first 380 pixels of vertical space, even visitors sad-
dled with small monitors can find what they seek without scrolling. Once
enmeshed in a story that engages their interest, visitors will scroll down a
few screens to continue reading.
How many screens of text will readers scroll before wearying of scrolling
and seeking the blessed release of clicking to the next page? Three. Just
kidding. Only a pseudo-scientist would pretend to know. As web designers,
we use our best judgment on each site. That, after all, is what we’re
getting paid for.
05 0732 CH03 4/24/01 11:16 AM Page 96
One reason frames are popular is that they allow web designers to keep the
interface onscreen in a consistent location, even when the user is scrolling
up or down like a madman. For instance, a horizontal menu bar at the top
(www.microsoft.com) or bottom (www.the-adstore.com) of the screen will
stay in place no matter how long the page may run and no matter how
much scrolling the user performs. Frames are on their way out (in W3C
parlance, they have been “deprecated”), but you can achieve the same
effects with CSS, a web standard.
Inexperienced designers sometimes create pages that require the user to

scroll horizontally. This is almost always unwise. Except at certain “art
gallery” sites, users will almost never scroll horizontally. Such interfaces are
inconvenient and often appear to be mistakes rather than deliberate design
decisions.
To understand why horizontal scrolling is an evil spawned from the fester-

ing loins of the incubus, imagine that you have to … turn the page to fin-
ish reading this … sentence and then fold the page back … to read the next
line of … text, which bleeds … backwards across the gatefold again, forc-
ing you to … turn the page, and then turn … it back again in order to begin
reading the next line.
No print designer would lay out book pages that way, but inexperienced
web designers do so frequently, whether from misguided creative impulses
or because they’ve made assumptions about their visitors’ monitor sizes.
This is another reason that Liquid Design (detailed in Chapter 2) comes
highly recommended; it always fits neatly into any user’s monitor.
It’s also the reason that clients, designers, and IT departments that set
“monitor baselines” of 800 x 600 are blockheads. If even 5% of the audi-
ence is expected to scroll horizontally simply to read marketing copy, the
client or web agency is effectively sending millions of potential customers
to a competitor’s site.
05 0732 CH03 4/24/01 11:16 AM Page 97
STOCK OPTIONS (PROVIDING ALTERNATIVES)

Users employ a variety of means to access the Web, including modern
browsers, older browsers, non-graphical browsers, audio browsers, and
non-traditional devices such as cell phones and PDAs. If the goal of a
site is to accommodate as many visitors as possible, then it is critical to
provide alternative forms of navigation.
Imagine that you have designed a lovely, frames-based site and that your
navigational menu exists in its own frame. A visitor using a text browser
enters the site. He cannot see frames because his browser does not sup-
port them. You, however, have thoughtfully included a <NOFRAMES> tag
in your HTML frameset. Inside the <NOFRAMES> tag you cut and paste the
main content from the home page, along with an HTML-based text menu.
The visitor can now use your content, even though he cannot see your
frames-based layout. (Again, we remind you that frames are on their way
out anyway.)
Options and alternatives increase the odds that someone will actually use
what you’ve designed. Larger web agencies employ quality assurance (QA)
staffs who spend all day hunting for online porn. Better QA staffers search
for flaws in your design by testing it in a wide variety of old and new
browsers on various platforms. Do not hate these site testers—they are your
friends. Build alternatives into your navigational scheme, and you will win
their admiration and more, importantly, that of your site’s audience.
The mechanics of including alternate forms of navigation will be covered

in Chapter 9, “Visual Tools.”
HIERARCHY AND THE SO-CALLED THREE

CLICK RULE
To accommodate the need for rapid access to information, a web designer
creates layouts that immediately reassure the visitor that she has “come
to the right place.” Brand-appropriate design accomplishes some of this
purpose. A clear hierarchical structure does the rest.
05 0732 CH03 4/24/01 11:16 AM Page 98
98 WHY: Where Am I? Navigation & Interface: Hierarchy and the So-Called Three Click Rule
It’s widely agreed, even by people who are not idiots, that web users are
driven by a desire for fast gratification. If they can’t find what they’re look-
ing for within three clicks, they might move on to somebody else’s site.
Hence the so-called “Three-Click Rule,” which, as you might expect, states
that users should ideally be able to reach their intended destination within
three mouse clicks.
With the average site offering hundreds if not thousands of items and
options, the Three-Click Rule sounds preposterous. But it is actually fairly
easy to achieve if you start by constructing user scenarios before you begin
to design the site.
What will people who use this site want to do? Where will they want to
go? Based on those scenarios, the site is structured into main areas of con-
tent. These are then organized into no more than five main areas. (See the
next section, “The So-Called Rule of Five.”) Submenus in each of the five
main areas get the user close enough that he or she is at least reassured
by the third click, even if it takes a fourth click to get to the final, desired
page.
Let’s play it out. You are designing a site for people who live with house-
cats. In the scenario portion of development, the team agrees that cat own-
ers might want to read about Mister Tibbles’ genetic heritage. In the
top-level hierarchy, you create an item called Breeds. When Aunt Martha
clicks Breeds, the site offers Long-Hair, Short-Hair, Tabby, and Exotic
options. A second click takes her to Short-Hair, a third to Mister Tibbles’
particular breed.
Like all so-called “laws” of web design, the Three-Click Rule is a sugges-
tion, not an ironclad rule. It is, though, a suggestion based on the way peo-
ple use the Web, and, particularly for informational and product sites, you
will find that it works more often than not. If nothing else, the rule can help
you create sites with intuitive, logical hierarchical structures—and that
ain’t bad.
05 0732 CH03 4/24/01 11:16 AM Page 99
THE SO-CALLED RULE OF FIVE

The so-called “Rule of Five” sounds like a period out of Chinese history, but
it’s actually just another guideline most working web designers keep in
mind—especially if they want to keep working.
The Rule of Five postulates that complex, multi-layered menus offering

more than five main choices tend to confuse web users. A glance back at
Figure 3.5 should confirm the common sense behind this “rule.” The main
menu at Overstock.com offers not five, not six, not seven, but a whopping
twelve main categories to choose from. (And that’s not even counting the
strange tagline area that is inexplicably designed to resemble a clickable
menu button.) Overstock.com is so busy offering everything that many
users will be hard pressed to find anything.
By contrast, Sapient’s main menu (back in Figure 3.14) offers four choices:
Clients, Expertise, Company Info, and Careers. Giving users three, four, or
five main choices makes it easier for them to decide where they want to
go. Hitting them with ten or more choices makes their next move harder
to predict—for them and for you. Confuse them enough, and it becomes
easier to predict where they will go, namely: anywhere else.
As with the Three-Click Rule, evolving a site whose architecture can be

navigated in five main areas or less is easier if you engage in scenario play-
ing before you begin to design. Chapter 7, “Riding the Project Life Cycle,”
provides a detailed analysis of how you, your team, and your client can col-
laborate to develop logical site structures that facilitate the Three-Click
Rule and the Rule of Five.
On multi-purpose sites (and there are many of those), several layers of nav-
igation may peacefully coexist. Looking yet again at Sapient (Figure 3.14),
four choices are enough to guide visitors to main areas of the site but not
enough to help those seeking one-click access to various client/vendor suc-
cess stories. The icon-driven menu on the right ignores the Rule of Five
without incident.
05 0732 CH03 4/24/01 11:16 AM Page 100
100 WHY: Where Am I? Navigation & Interface: The So-Called Rule of Five
On a shopping site, the main menu may offer three choices: Women’s,
Men’s, and Kids’. But submenus can be far more extensive: the Women’s
section might offer Outerwear, Sportswear, Business Attire, Casual Wear,
Accessories, Cosmetics, Health Aids, and sundry other stuff without con-
fusing any shopper. As the shopper burrows deeper into the hierarchy, these
submenus can sprout submenus of their own, for example Cosmetics could
include Hair Products, Makeup, Toners, Cleansers, and beyond. Such sub-
menus may run deep, as long as they appear when users expect them to
appear and behave consistently from section to section.
Some site designers and architects distinguish between goal- and task-ori-
ented navigation. With goal-oriented navigation, the user wants to go
somewhere (Clients, Expertise, or Company Info, for example). With task-
oriented navigation, the user wants to do something (apply for a job, log
in, or read case studies). Combining the two types of user needs in the same
navigational context can be more confusing than helpful. In such cases,
task and goal-oriented navigation coexist separately (see Figure 3.16), and
the Rule of Five pertains to each navigational stream rather than to the
page as a whole.
Figure 3.16
Goal-oriented navigation
(Expertise, Process, Proof)
and task-oriented naviga-
tion (Hire Us, Work Here,
Login) carefully separated
and balanced. The user
can quickly follow a
desired activity path
without becoming con-
fused or overwhelmed.
Such complex structures
are hard to pull off
(www.hesketh.com).
05 0732 CH03 4/24/01 11:16 AM Page 101
HIGHLIGHTS AND BREADCRUMBS

Drivers use road signs to track their location in space. Web users rely on
navigation. Well-designed sites cue the visitor to her location within the
site’s hierarchy. For instance, if the visitor is within the Breeds section of
the cat site, the Breeds item in the menu bar may be highlighted by a sub-
tle change of color. This “you are here” indicator helps keep the visitor
grounded, thus promoting lengthier visits (see Figure 3.17).
Figure 3.17
Subtle highlighting on the
menu bar reminds you that
you’re on the Home Store
page. Click to a different
page, and a different menu
item will be highlighted.
Note, too, how much air
the design team has
managed to work into the
page, in spite of the vast
number of links and menu
items the page must carry
(www.bloomingdales.com).
Compare with Figure 3.16
and contrast with Figure 3.5.
It’s all about comfort. Better hotels offer fluffier pillows; better sites pro-
vide constant spatial and hierarchical reassurance. Breadcrumbs, called
this because they resemble the trails left by Hansel and Gretel, not only
serve as hierarchical location finders, but they also allow visitors to jump
to any section further up in the hierarchy (see Figure 3.18).
05 0732 CH03 4/24/01 11:16 AM Page 102
102 WHY: Where Am I? Navigation & Interface: Consistent Placement
Figure 3.18
Breadcrumbs remind you
that you’re on the Miles
Davis page of the Artists
section. Essential to
complex directories,
breadcrumbs can enhance
branding, entertainment,
and content sites by pro-
viding alternative naviga-
tion for those using
less-capable browsers.
They reassure beginners
while enabling sophisti-
cated users to skip tedious
hierarchical layers and
move quickly to the exact
content they seek
CONSISTENT PLACEMENT
The location of the navigation in the digital nation permits much permu-
tation without causing perturbation. Navigation can exist in a horizontal
strip at the top or bottom of the site. It can live in a navigation bar on the
left or right side of the page. It also can float in a JavaScript remote popup
window (as long as alternatives are provided).
What matters most, aside from technological and user appropriateness

(remote popup window navigation is probably not the best choice for the
Happy Valley Retirement Home), is that the navigation stay in one place so
the user knows where to find it when he or she is ready to move on. A
handrail guides someone down a flight of stairs, and the guidance works
because the handrail remains in the location where the user expects to find
it. Good site navigation works the same way. With few exceptions, it does-
n’t really matter where you stick your navigation as long as you keep stick-
ing it there throughout the site.
05 0732 CH03 4/24/01 11:16 AM Page 103
Figure 3.19
What’s the “best” place
for navigational menus?
That’s up to the web
designer. Caffe Mocha
runs its menu bar
horizontally across the
middle of the page
(www.caffemocha.com).
BRAND THAT SUCKER!

We’ve discussed navigation and interface in terms of the user’s needs, and
they of course come first. But what of the client’s needs? Meeting them is
the role of branding.
A corporate website is the online expression of that company’s brand iden-

tity. Making sure that the navigation fully supports the company’s brand
identity is crucial to the success of the site (and sometimes to the success
of that company). Build the most navigable, information-filled site in the
world, and if it lacks a coherent brand identity, nobody will remember it,
nobody will tell their friends about it, and nobody will bother to bookmark
it and return.
For over 100 years, advertisers have been working to build our joyful world
of branding. When your stomach hurts, you reach for Alka-Seltzer (not an
antacid). Sneeze, and you reach for Kleenex (not a disposable paper tissue).
05 0732 CH03 4/24/01 11:16 AM Page 104
104 WHY: Where Am I? Navigation & Interface: Brand That Sucker!
Like millions, we may express our individuality through Levi’s. You may
choose Gap to show the world how different you are. Neither of us, as we
don our separate uniforms, is likely thinking about the folks who picked the
cotton, or groomed the silkworms, or trimmed the fleece from the sheep.
Consciously or unconsciously, we’re identifying ourselves with images cre-
ated in small offices, thousands of miles from where the cotton grows and
the silkworm arches toward the sun—images created by brand advertising.
Branding, branding, branding. McDonald’s does not sell cereal mixed with
the flesh of cows; it sells food, folks, and fun. Marlboro sells the myth of
the freedom of the Wild West. Camels are not for everybody, but then, they
don’t try to be.
Branding is not limited to products. Although his verbal gymnastics, half-

spoken vocal delivery, and angry social consciousness predate Rap, Bob
Dylan can’t perform Hip Hop; it would conflict with his brand image as the
spokesman of the 1960s generation. But David Bowie can do hip-hop or
drums-and-bass because his brand identity is that of an ever-changing,
ever-current chameleon.
And how come Seinfeld can quip wisecracks about supermarket checkout
lines but will never mine his personal sexual experiences for comic mate-
rial? Hey, it’s not part of the brand.
How does this relate to the task of web design? As a designer, you know
the answer to this one already. Whether you’re building a corporate site or
a multimedia online funhouse, your first task is to understand and trans-
late the existing brand to the web medium or to create a new brand from
scratch.
Good interfaces reflect the brand. Sleek, high-tech graphics complement a

sleek, high-tech company—or one that wants to be perceived that way. A
“friendly” GUI is necessary for a “friendly” company such as AOL. (You in
the back, keep your sarcastic observations to yourself.) It goes without say-
ing that the company’s color scheme, logo, and typographic style must be
reflected in your web graphics and that existing print and other materials
are often a guideline to what is appropriate for the site.
05 0732 CH03 4/24/01 11:16 AM Page 105
Smart web designers go far beyond the obvious. In addition to graphic

design elements, savvy web folk craft interfaces whose very functioning
reflects and extends the brand. A “fun” brand needs more than cute graph-
ics. Its sectional titles should be fun to read and its menu fun to interact
with. This may mean taking a cue from the world of gaming. It may mean
building the interface in Macromedia Flash.
A movie studio’s interface should not resemble that of a bank. A company

that sells active wear should encourage active participation, through
games, message boards, or contests. A literary site’s interface should qui-
etly promote reading, instead of busily distracting from it with funky danc-
ing icons. (A literary site that avoids long copy belies its own brand
identity.) The interface of a religious organization’s site dare not resemble
that of an e-commerce site, lest visitors along with moneylenders be driven
from the temple.
IBM’s brand image is that of a big-time solutions provider (www.ibm.com).

If you’re asked to design their site, it had better be technologically solid,
visually impeccable, and easy to use. Anything less will send the wrong
brand message.
■ Technologically solid, in this brand context, doesn’t mean a deluge

of plug-ins or a reliance on safe, old 1990s web technologies; it
means online forms that work, search functions that deliver usable
results, and enhancements that shine in new browsers while degrad-
ing well in old ones.
■ Visually impeccable means that imagery and typographic choices

must play in the key of the brand. Type should be clean and read-
able—not fussy, not grungy, not softly feminine or boyishly abrasive.
Photographic images need not be disgustingly corporate (two suits
at a monitor will take you only so far), but images of crime, drugs, or
bongo jams will obviously be inappropriate.
■ Easy to use means easy to use. Why even mention it? Because if vis-
itors find their way to content they seek on the IBM site, it reinforces
the overriding brand idea that IBM provides solutions. If users get lost
or don’t know which button to push, it will send the opposite mes-
sage. Sending the wrong brand message could harm a brand identity
the company has carefully built up over generations.
05 0732 CH03 4/24/01 11:16 AM Page 106
106 WHY: Where Am I? Navigation & Interface: Brand That Sucker!
Branding the WaSP
The Web Standards Project (WaSP), mentioned in Chapter 2, evolved from con-
versations between a number of frustrated web designers and developers.
While some members brought high-level technological knowledge to the proj-
ect and others brought “marquee value” (their names alone adding instant
credibility to anything the WaSP might say or do), your humble author focused
on creating a brand identity that would be both memorable and consistent
with the task at hand.
Many names were bandied about; we pushed “The Web Standards Project” for
a variety of reasons, not least of which was its ability to be referred to in short-
hand by the acronym WaSP. Call us shallow, but we believed that this aggres-
sive little insect was the perfect metaphor for our group. We also knew that
a memorable identity was needed to keep the effort from becoming so tech-
nologically-focused as to confuse potential members.
After all, by agitating for compliance with web standards, we were taking on
giant companies such as Netscape and Microsoft in spite of being a small
grassroots effort. Which tiny creature has the power to disturb a giant? The
wasp. It’s a purposeful, productive beast with a powerful stinger, and while
you may be able to swat away one wasp, you don’t want to mess with an angry
nest. The site’s verbal tone and visual approach came straight out of this sim-
ple little brand image—from the color palette (wasp-yellow, gold, and black)
to the tone of voice (www.webstandards.org).
When Kioken Inc. (www.kioken.com), a leading New York web shop, was
charged with designing a site for the high-end retailer, Barney’s, they
carefully considered the client’s brand identity as a provider of well-made,
tasteful, and luxurious clothing. To put it bluntly, Barney’s goods are well
above the means of most of us working stiffs, and Barney’s customers like
it that way.
Kioken crafted a sophisticated, Flash-based interface like nothing else on

the Web (www.barneys.com). If you were a savvy web user, owned a fairly
powerful PC, had a fast connection, and were equipped with the latest
Flash plug-in, you were treated to a unique showcase of Barney’s clothing.
Just navigating it made you feel smarter than the average web user.
If you were not an experienced web user, owned an old PC, had not down-
loaded the latest Flash plug-in, and were stuck with a slow dialup modem
connection, Kioken (and their client) figured that you were not really a Bar-
05 0732 CH03 4/24/01 11:16 AM Page 107
ney’s customer anyway. A certain elitism was as much as part of the inter-
face as it is of the store. The Barney’s site may not exemplify democratic
humanism, but it is a perfect web translation of the client’s brand.
Some critics faulted Barneys.com for failing to provide an e-commerce

solution. You could look at Barney’s clothing, but you could not buy it
online. The criticism betrays a misunderstanding of the client’s brand iden-
tity. You expect to be able to buy jeans from Sears’ website, but to buy Bar-
ney’s clothing online would be wrong for such a highfalutin’ brand.
Interfaces that deeply and meaningfully reflect the brand will encourage
repeat user visits and repeat assignments from your clients. As a web
designer grounded in traditional art direction and design, you are better
equipped than many working professionals to create brand-appropriate
web interfaces: interfaces that don’t just look like the brand, they behave
like it.
Interfaces that look and act like the brand and that guide the right audi-
ence to the most important content or transactions form the foundation
for the best sites on the Web—the ones you are about to design.
05 0732 CH03 4/24/01 11:16 AM Page 108
06 0732 Part II 4/24/01 11:17 AM Page 109
Part II
WHO: People, Parts,

and Processes
4 How This Web Thing Got Started 111
5 The Obligatory Glossary 123
6 What Is a Web Designer, Anyway? 135
7 Riding the Project Life Cycle 147
06 0732 Part II 4/24/01 11:17 AM Page 110
07 0732 CH04 4/24/01 1:02 PM Page 111
chapter 4
How This Web Thing Got

Started
1452
GUTENBERG CONCEIVES OF MOVEABLE TYPE based on a punch-and-mould
system. Working with paper (brought to Europe from China in the twelfth
century), oil-based ink, block print (brought to Europe by Marco Polo in the
thirteenth century) and a wine press, he sets the stage for the mass pro-
duction of books and the wide dissemination of learning.
1836
Cooke and Wheatstone patent the telegraph, thus bringing telecommuni-
cations to the world. For the first time in history, two people can carry on
an argument even when they are miles apart.
1858
The first Atlantic cable is laid across the ocean floor, facilitating telecom-
munications between Europe and the United States. Unfortunately, the
cable goes on the fritz after just a few days. (And you thought your cable
service was bad.) A second attempt in 1866 succeeds. That cable will
remain in service for close to a century.
07 0732 CH04 4/24/01 1:02 PM Page 112
112 WHO: How This Web Thing Got Started: Why We Mentioned These Things
1876
Alexander Graham Bell demonstrates the telephone. The first busy signal
follows soon after.
WHY WE MENTIONED THESE THINGS

The events we just mentioned set the stage for the Internet and thus even-
tually for the Web. Gutenberg’s invention sets in motion the concept that
information belongs to the people (at least, to those people with a few
coins in the pockets of their funny fifteenth-century pants). The subse-
quent technological breakthroughs make possible the eventual sharing of
data via telephone lines.
1945
Vannevar Bush, Science Advisor to U.S. President Roosevelt, proposes
a “conceptual machine” that can store vast amounts of information
linked by user-created associations. He calls these user-generated con-
nections “trails and associations.” Eventually they’ll be called “hyper-
links.” (As We May Think, www.theatlantic.com/unbound/flashbks/
computer/bushf.htm).
1962
The Advanced Research Projects Agency Network (ARPANET) is established,
which will eventually be known as the Internet. Dr. J.C.R. Licklider is
assigned to lead ARPA’s research into the military application of computer
technology.
1965
Scientist Ted Nelson coins the word hypertext to describe
“nonsequential writing—text that branches and allows choice
to the reader, best read at an interactive screen.” (See http://
www.acclarke.co.uk/1960-1969.html and http://ei.cs.vt.edu/
~wwwbtb/book/chap1/htx_hist.html for more information.)
07 0732 CH04 4/24/01 1:02 PM Page 113
Nelson dreams of a worldwide library of all human knowledge that can be

read on a screen and based on links. Sound familiar? Nelson also dreams
of micropayment-based royalty schemes, two-way links, and other fea-
tures not found in the Web as we know it.
1966
ARPA scientist Robert Taylor, no doubt depressed when women find out he
is not the movie star Robert Taylor, figures out a way for researchers at var-
ious locations to collaborate by means of electronic computer networks.
Inexpensive terminals are linked to a few pricey mainframe computers. Sci-
entists begin exchanging documents and email messages. The first public
demonstration of what is now being called ARPANET will take place in
1972. The Internet is born.
1978
On January 3, Steve Jobs and friend, Woz, take Apple Computer public, thus
launching the personal computer “revolution.” As Gutenberg’s invention
brought human knowledge out of the monastery and into the hands of
ordinary citizens, Jobs and Woz’s invention takes the arcane business of
data crunching out of the realm of Big Science and makes it available to
folks like us. The subsequent Macintosh computer (1984) offers a Graphi-
cal User Interface (GUI), making it easier still for ordinary people to use a
computer. The Graphical User Interface, based on work done in Xerox Parc
in the 1970s, enables people to perform tasks by clicking onscreen icons
and buttons. Most civilians find this easier than memorizing and typing
cryptic commands. A Windows GUI follows in the PC realm. The point-and-
click interface will be key to the eventual acceptance of the Web.
1981
The domain name server (DNS) is developed, thus making the future
safe for web addresses (www.ietf.org/rfc/rfc0799.txt). At first these will
have cryptic numerical “names” such as 191.37.4211, but eventually
consumer-friendly domain names such as brandname.com will take their
place. This is key because advertisers would see little value in adding
“Visit us at 191.37.4211” to the end of their radio commercials but are
07 0732 CH04 4/24/01 1:02 PM Page 114
perfectly happy asking us to visit brandname.com. When advertisers are

happy, they spend money. When money is available, professionals
arise to claim it. The rise of web design and development is thus
partially made possible by the invention of consumer-friendly domain
names.
1984
The Apple Macintosh ushers in an era of “desktop publishing,” empowering
designers to set their own type and place and color-correct their own
images, rather than relying on the skills of third-party service profession-
als. Desktop publishing also empowers ordinary citizens to express them-
selves creatively, sometimes (though not always) with wonderful results.
This too will be mirrored a decade later, when the Web empowers anyone
with a computer and the willingness to learn HTML to become a “web
designer.”
As if all that was not enough, Apple makes use of Ted Nelson’s hypertex
concept in its HyperCard product, which enables creative folks to create
link-based presentations.
1986
There are now 5,000 Internet hosts (computers connected to the Internet
“backbone”) and 241 newsgroups.
On the campaign trail, Al Gore makes frequent reference to the developing

“Information Superhighway.” The phrase actually refers to high-speed
coaxial networks, but it is popularly understood to mean ARPANET or the
Internet. Press confusion on the subject will later haunt Gore’s 2000 bid for
the U.S. presidency.
1988
The NSFNET backbone is upgraded to T1 (1.544 Mbps). We’re not sure what
this means either, except that stuff gets a lot faster.
Internet Relay Chat (IRC) is developed in Israel, thus paving the way for
a future where office workers can complain about their jobs to friends
in foreign lands, instead of simply boring their spouses with these petty
grievances.
07 0732 CH04 4/24/01 1:02 PM Page 115
1989
Tim Bray and others cofound Open Text, an Internet search engine. Search
engines cut through the chaos of the burgeoning Internet by enabling cit-
izens to actually find things. This ability to find things brings value to the
Net and will be an invaluable aspect of the coming Web. Search engines
will eventually enable citizens to find half-price airline tickets or seek out
information to help their children write school reports. The human and
commercial potential built into that premise will empower the coming
“revolution” of faster and faster networks, and larger and larger web agen-
cies such as Scient, iXL, and Razorfish.
CERN is the biggest Internet site (location) in Europe. Working there is a

young scientist, Tim Berners-Lee.
1990
On the twelfth of November at CERN, Tim Berners-Lee (with R.
Cailliau) invents the World Wide Web, rooting the idea in hypertext:
“HyperText is a way to link and access information of various kinds as a

web of nodes in which the user can browse at will… A program which pro-
vides access to the hypertext world we call a browser… World Wide Web
(or W3) intends to cater for these services across the HEP [High Energy
Physics] community.” (See http://www.w3.org/Proposal.)
Not content with the profundity of this invention, Berners-Lee also devel-
ops a “web browser” on his NeXT machine. With Berners-Lee’s browser, not
only can you view web pages, you can also edit and design them. Fortu-
nately, the “designing” part of the browser does not make it far out of
Berners-Lee’s lab, and thus the way is paved for professional designers and
art directors, rather than scientists, to create the visual language of the
Web. (The original CERN W3 package included a server, a browser, and a
true WYSIWYG editor.)
1991
America Online (AOL) begins offering Internet access in addition to its pro-
prietary content and newsgroup features. Millions of people begin “going
online” thanks to AOL’s easy-to-use point-and-click functionality and con-
sumer-friendly brand imagery. This is important because if the Internet had
07 0732 CH04 4/24/01 1:02 PM Page 116
remained the province of geeks, the Web would not have gained such ready
acceptance, let alone exploded into public consciousness. You would not
be thinking about a career in web design, and this book would be all about
delicious low-fat recipes rather than the Web.
1993
January: Marc Andreessen and Eric Bina, young programmers working for
the National Center for Supercomputing Applications (NCSA) invent a
point-and-click graphical browser for the Web, designed to run on UNIX
machines. It is called Mosaic because the name Pantaloons didn’t do as
well in testing. (Just kidding. Not kidding about Mosaic, they did indeed call
it that. Just kidding about why they called it that because we frankly don’t
know and this paragraph felt a little “short” to us.)
August: Andreessen and his co-workers release free versions of Mosaic for
Macintosh and Windows PCs.
December: Andreessen quits his day job.
There are two million Internet hosts and 600 websites.
The NCSA “What’s New” page (www.ncsa.uiuc.edu/SDG/Software/Mosaic/

Docs/whats-new.html) is both an early non-commercial web directory
and one of the first weblogs. A weblog is a frequently updated, annotated
directory of stuff on the Web. In 1998, weblogs (always quietly pres-
ent) would “catch on” again thanks to sites such as Scripting News
(scripting.com), Robot Wisdom (www.robotwisdom.com), and Memepool
(www.memepool.com). By 1999 they would become downright trendy, as
hundreds of web designers create personal weblogs to keep their friends
abreast of the sites they like, while thousands of first-time web publishers
use tools such as Blogger, Manila, and Pitas to produce their own personal
“Blogs.”
1994
Marc Andreessen hooks up with Jim Clark, founder of Silicon Graphics Inc.
The two form a company called Mosaic Communications Corporation to
promote their Netscape web browser. NCSA, holders of the Mosaic trade-
mark, balk at this use of their trademark, eventually prompting the young
browser company to rename itself Netscape Communications.
07 0732 CH04 4/24/01 1:02 PM Page 117
Two graduate students, Jerry Yang and David Filo, form Yahoo! (Yet Another
Hierarchical Officious Oracle), a directory whose purpose is to keep track
of the websites springing up everywhere (www.yahoo.com). The site is
organized somewhat like a library’s card catalog system. Other directories
of lesser quality quickly spring up in imitation.
Wired Magazine’s Hotwired site evangelizes the new medium and pioneers
techniques of web design and web architecture.
Tim Berners-Lee founds the World Wide Web Consortium (W3C), an inter-
national non-profit think tank dedicated to providing a rational roadmap
for the technological advancement of the Web.
People begin designing and producing personal sites because they can.
“Justin’s Links from the Underground” (www.links.com) is one of the fir-
st and most famous personal sites. Glenn Davis launches Cool Site of
the Day (www.coolsiteoftheday.com) to keep track of interesting or funky
content on the rapidly growing Web.
1995
Pushed into public consciousness and acceptance by the coolness of
Netscape’s Navigator graphical browser and by sites such as Cool Site of
the Day, the Web mushrooms. There are now 6.5 million hosts and 100,000
websites.
The Web functions well, but its design potential is sadly underdeveloped.
David Siegel, a typographer and early web designer, publishes “Web Wonk”
(www.dsiegel.com/tips/), an online tutorial offering techniques with which
designers can create pleasing, magazine-like page layouts on the Web by
working around (hacking) the limitations of HTML—the language with
which web pages are created. These techniques seriously conflict with the
purpose of HTML as a simple, structured language for sharing documents.
But they are all designers have to work with at this time. The rift between
the W3C and graphic designers has begun. (In 1996, Siegel publishes the
book, Creating Killer Websites. Though far from the first how-to guide, it
will be one of the first books to treat web design as a serious issue.)
Netscape introduces the tiled background image in Navigator 1.1. Warner

Brothers’ “Batman Forever” site is among the first to make intelligent use
of the feature, hacking it to create the illusion of full-screen images.
07 0732 CH04 4/24/01 1:02 PM Page 118
Batmanforever.com helps prove that the Web has tremendous potential for
anyone wishing to promote an idea, event, or product. There are three mil-
lion web users, and half of them—1.5 million people—view this one site
every week.
Jakob Nielsen, a Ph.D. from Sun Microsystems, begins publishing articles

(www.useit.com) calling for a rational approach to the development of the
Web. Nielsen calls his approach “usability” and claims that it is based on
scientific studies. The rift between designers and usability experts has
begun.
Personal home pages are proliferating.
Yahoo! and other large sites begin running ad banners.
Netscape goes public.
1996
David Siegel creates “High Five” to honor and showcase the
best-designed sites on the Web. (High Five is no longer active, but
archives are available at highfivearchive.com/core/index.html.) He bestows
the first High Five award on his own site. Some consider the gesture arro-
gant, but Siegel doesn’t care; his book is selling like crack. And, to some
extent because of his evangelism, the Web begins attracting greater num-
bers of design professionals and becoming better and better designed as a
result. But this aesthetic boon comes at a cost. Because most of us are
using hacks and workarounds to make our sites more attractive and read-
able, few of us are demanding the creation of robust standards that would
provide better presentational capabilities without breaking the Web’s
structural underpinnings. And since we’re not hollering for better stan-
dards, the W3C isn’t rushing them out the door, and browser makers aren’t
hastening to support them. We will all pay for this later.
“Suck” (suck.com), a brilliantly written daily site created by Joey Anuff and
Carl Steadman, offers sardonic commentary along with a radically flat-
tened hierarchy. Instead of offering a splash page, followed by a contents
page, followed by sectional header pages, and so on (the tedious architec-
ture found in most early sites), Suck slaps its content on the front page
07 0732 CH04 4/24/01 1:02 PM Page 119
where you can’t miss it. Minds reel. The rift between web architects and
graphic designers begins. (Architects think about streamlining and con-
trolling the flow of the user’s experience. Graphic designers think about
reinventing the interface and blowing the user away on every page. Good
web designers struggle to find a balance between these two approaches on
a site-by-site basis.)
Anuff and Steadman will later sell their creation to their employers for
more than lunch money, thus ushering in a period where “content is king,”
whether it’s actually valuable or even read, and where everybody and her
sister wants to be a millionaire. This is not Anuff or Steadman’s fault.
Word.com begins offering intricately designed, well-written content. Like

Suck, Word.com will be purchased later, with mixed results. One mass delu-
sion (“content is dead”) will briefly replace another (“we all get to be mil-
lionaires”).
Netscape introduces JavaScript, a “simple” programming language that

enables web pages to become far more interactive. Web designers begin
stealing JavaScript from each other.
Netscape and Sun announce that Sun’s new object-oriented Java language
will “free” everyone from the “tyranny” of Microsoft’s Windows operating
system. Bill Gates smells the coffee. Microsoft creates Internet Explorer.
The browser wars begin. Over the next four years, Netscape will invent one
way of doing things while Microsoft invents another. Web designers will be
forced to choose which technologies to support—or will support multiple
technologies at considerable cost to their clients. Eventually, most every-
one will realize that the medium can only advance with full support for
common standards.
There are 12.8 million hosts and half a million websites.
1997
Amazon.com begins selling books over the Web. Marketers everywhere
wake up to the promise of e-commerce and begin scrambling to launch e-
commerce companies, add e-commerce capabilities to the offerings of
their existing companies, or just put the letter “e” in front of whatever it is
that they do. There are e-books, e-investments, e-architects, and e-com-
munities. E-nough, already. A brief i-period will follow the e-period.
07 0732 CH04 4/24/01 1:02 PM Page 120
Internet Explorer 3.0 begins to support Cascading Style Sheets (CSS), an

advanced yet simple-to-use design technology created by the W3C.
Netscape Navigator 3.0 does not support CSS but does offer JavaScript
(and JavaScript Style Sheets—a competing technology that nobody ever
adopts). IE3 does not fully support JavaScript. The browser wars escalate,
and the Web becomes still more fragmented.
There are now 19.5 million hosts, one million websites, and 71,618 news-
groups.
1998
There are over 300 million pages on the Web—and 1.5 million new ones
appear online daily.
Internet traffic doubles every 100 days.
Investors become frenzied. Venture capitalists become stupidly wealthy.

Anyone in a suit can raise $5 million by promising to sell anything to any-
body. If we exaggerate, it’s because this is a period of deep delusional
dementia fueled by 80s style greed and 90s style buzzwords. Baby Jesus
weeps.
The growth of e-commerce exceeds its one-year expectation by more than

10,000 percent. The projected growth of business-to-business services on
the Web dwarfs even the growth of e-commerce.
With much money at stake, the browser war’s fragmentation of the Web
becomes intolerable. Developers spend at least 25 percent of their time
working around incompatibilities between Netscape and Microsoft
browsers.
A group of designers, developers, and writers, lead by Glenn Davis

and George Olsen, forms The Web Standards Project (WaSP) at
www.webstandards.org. The group hopes to persuade browser makers to
support common standards so the Web can evolve rationally.
The W3C, which creates most of the standards, lacks police power
to enforce them; in W3C parlance, things such as CSS and HTML 4
are “recommendations.” The WaSP sees these recommendations as an
absolute necessity and will spend the next three years spreading that
gospel by any means necessary.
07 0732 CH04 4/24/01 1:02 PM Page 121
Netscape goes open source, unveiling the secrets of its code in the hopes
that thousands of programmers around the world will join together to cre-
ate a newer, better version of the Netscape browser. The open source proj-
ect for the Netscape Navigator source code is named Mozilla. The
Department of Justice begins an antitrust lawsuit against Microsoft.
1999
America Online (AOL), though partially responsible for the growth and pop-
ularity of the Web, has long been despised by Internet connoisseurs. Many
holding this view are die-hard Netscape users, who see AOL as a propri-
etary service for frightened “newbies” (neophyte Internet users). In a move
that shocks the online world, AOL buys Netscape.
Netscape announces that its upcoming 5.0 browser, being built by

the Mozilla open source project, will fully support the five key stan-
dards demanded by The Web Standards Project (www.webstandards.org
/mission.html). The 5.0 browser never sees the light of day, but in late 2000
the project and Netscape will give birth to Netscape Navigator 6.
Microsoft announces that its upcoming 5.0 browser for the Mac will fully
support two key web standards and offer “90 percent support” for others.
At least 100,000 web-related jobs cannot be filled because of lack of qual-

ified personnel. Populi, the Web Talent Incubator, is launched to solve this
problem. Your humble web author, who appears to enjoy typing the phrase
“your humble web author,” will later help Populi develop a curriculum in
web communication design, which will still later become the basis for the
book you are now reading, which will yet later be unearthed by archeolo-
gists of the thirty-first century, along with a Pepsi bottle.
2000
The year web standards broke, 1
Internet Explorer 5, Macintosh Edition is released in March, offering
near perfect support for HTML 4, CSS-1, and JavaScript
(www.alistapart.com/stories/ie5mac/).
07 0732 CH04 4/24/01 1:02 PM Page 122

Netscape 6 is released in the wee hours of November 14. It supports XML,
and the W3C DOM as well as the standards supported by IE5/Mac.

Opera 5 (www.opera.com), released in December, supports HTML, CSS, XML,
WML, ECMAScript, and the DOM (www.opera.com/opera5/specs.html).
The year the bubble burst

A number of ill-conceived web businesses fail, causing the usual dire
predictions and market panics. A number of good web businesses
are dragged down along with the unworthy ones. Overbuilt web agencies
lay off staff; other agencies absorb them.
2001
You buy this book. And buy a second copy for a friend. And a third for your
coffee table.
08 0732 CH05 4/24/01 11:18 AM Page 123
chapter 5
The Obligatory Glossary
SEVERAL YEARS BACK, Grey Advertising, Inc. felt it was perceived as a some-
what lackluster agency: large, dependable, and successful at delivering
results, but not exactly cutting-edge in a world of Chiats and Weiden-
Kennedys (the people who have made commercials for Apple and Nike).
Grey wanted to enhance its image, and as companies often do, it brought
in an outside consultant. A depressing sum of money later, the consultant
unveiled this recommendation: make the logo orange. A Grey company
with an orange logo, get it? Unexpected. Cutting edge. Fresh. Or so the con-
sultant argued, and the agency apparently agreed.
The story may be apocryphal, we hasten to add, because Grey has more
lawyers than our publisher. We mention the whole thing because, as if
Internet terminology itself weren’t confusing enough, job nomenclature at
web agencies can be dazzlingly baffling. This is thanks, in part, to consult-
ants who think an orange Grey makes an Apple and “user experience trans-
actional information architect” sounds better than “designer.”
The Web is an insanely great medium. The young industry is exciting

and challenging enough to fulfill you through a dozen lifetimes, but the
business is so new that even people who work in it get confused over
terminology.
08 0732 CH05 4/24/01 11:18 AM Page 124
124 WHO: The Obligatory Glossary: Web Lingo
Some companies have a dozen different titles for designers with slightly
different jobs; other companies slap one title on everybody, and often
enough the title makes little intuitive sense. Orange you Grey we’ve pro-
vided this little chapter to help you navigate the twin minefields of Inter-
net buzzwords and ever-changing job titles? You bet you are. (Our
apologies to Grey Advertising, consultants everywhere, and People for the
Ethical Treatment of Animals, whom we haven’t offended but just felt like
mentioning because it’s a good cause. Besides, if we don’t mention it here,
our cats will claw our eyes out—and they can do it.)
WEB LINGO
Extranet
An extranet is a private network of computers that is created by connect-
ing two or more intranets or by exposing an intranet to specific external
users and no one else. Business-to-business collaboration often uses
extranets.
In English: Extranets are websites that allow Company A to interact with

Company B, and Special Customer C to interact with either or both—pretty
kinky stuff. As a web designer, you may never be called upon to design an
extranet. (If you are, it’s the same thing as designing a website. We’re sorry
to bore you with these tedious distinctions, but that’s our job in a section
like this. We hear the American Movie Classics cable network is hosting an
Alfred Hitchcock retrospective. Maybe you should go watch it until this
chapter blows over.)
On the other hand, the Business-to-Business (B2B) category is one of the

largest growth areas of the Web, so you may find yourself stuck, er, asked
to design extranet sites anyway.
Websites are websites whether they’re designed for the general public or
for private businesses. However, because extranets are business-oriented,
they tend to be more like software and less like magazines or television. In
other words, the challenges are closer to industrial design and technical
design and further from the consumer-oriented design many of us are used
08 0732 CH05 4/24/01 11:18 AM Page 125
to. In still other words, this type of design work is not for everybody, though
some designers adore and excel at it. (Excel is a trademark of Microsoft,
and even though we didn’t use it in that context in the preceding sentence,
their lawyers read everything.)
HTML
Hypertext Markup Language (HTML) is an application of Standard Gener-
alized Markup Language (SGML) and is used to construct hypertext docu-
ments (web pages).
In English: HTML is to web pages what PostScript is to print. But while Post-
Script is a complex programming language, best handled behind the scenes
by software such as Illustrator and Quark XPress, HTML is a simple markup
language best written by human beings. HTML breaks content down into
structural components, much as an outline does.
The simplicity of HTML makes it easy to learn, but that simplicity also can
be limiting. Soon, many sites will be built with more advanced tools, such
as Extensible Markup Language (XML). You need not concern yourself with
that now. Later on in this book we will show you what HTML is, how to use
it correctly, and how to employ it creatively. See Chapter 8, “HTML: The
Building Blocks of Life Itself.”
Hypertext, hyperlinks, and links

For additional information, refer to the section titled, “Website” later in this
chapter.
Internet
The Internet is a worldwide networking infrastructure that connects all
variety of computers together. These connections are made via Internet
protocols including (surprise, surprise) Internet Protocol (IP), Transport
Control Protocol (TCP), and User Datagram Protocol (UDP). IP is used for
addresses, TCP is used to manage sockets (and hence the Web), and UDP is
used to manage Domain Name Servers (DNSs). See Chapter 4, “How This
Web Thing Got Started,” for further explanation.
08 0732 CH05 4/24/01 11:18 AM Page 126
In English: The Internet is to the Web as cable networks are to television

or as phone cables and switching stations are to your Uncle Marvin, who
always phones while you’re away on vacation and then resents you for not
returning his call the very next day. The Internet is a combination of hard-
ware (computers linked together) and software (languages and protocols
that make the whole thing work).
As a web designer, all you need to keep in mind is that you’re not only com-
municating with readers and viewers (“users” if you must), you’re also cre-
ating work that must fit into formats appropriate to Internet technology.
In other words, it’s not your job to manage networks (for instance) as long
as you understand their implications for your work—such as bandwidth and
cross-platform issues. See Chapter 2, “Designing for the Medium.”
Intranet
An intranet is an internal or private networking infrastructure that uses
Internet technologies and tools. Unlike what occurs on the Internet, only
computers on the private physical network can access an intranet.
In English: As a web designer, in addition to creating sites for the public,

you also might be called upon to create intranet sites, which are nothing
more than websites for private companies. For instance, AT&T not only has
websites for the public, it also has thousands of private intranet sites where
its employees can communicate with each other, schedule appointments,
keep track of company policies, and so on.
One other difference worth noting is that when you’re designing an Inter-
net site, it has to be usable by anyone in the world—Netscape, Opera, IE,
and iCab users; 6.0 browser users as well as 2.0 browser users; the blind
and the not-blind; WebTV users and AOL users alike. You get the picture.
On an intranet site, by contrast, all visitors may be using the same web
browser and computing platform, which can simplify some of your design
choices.
Of course, even in such circumstances, it is best to design with open stan-

dards so that your client will not be locked into restrictive choices. For
instance, if you had designed an intranet for a network of Netscape 4 users
08 0732 CH05 4/24/01 11:18 AM Page 127
in 1998, you might have built the entire site using Netscape’s proprietary
LAYERS technology. But with Navigator 6, Netscape stopped supporting
LAYERS in favor of W3C standards. Had you designed specifically for
Netscape’s previous browser, your site would not work when the client
upgraded browsers. Clients dislike that sort of thing, even when they are
the ones who insisted on using a specific technology. Proceed with caution.
Additionally, if all the site’s users are connected via a local network, you
can make bold use of bandwidth-intensive technologies such as streaming
video. When designing for the Web, you need to worry about bandwidth.
Full-screen video is out; smaller video images and heavily compressed
audio might be okay. For more on this fascinating topic, see Dave Linabury’s
“The Ins and Outs of Intranets” at www.alistapart.com/stories/inout/.
JavaScript, ECMAScript, CSS, XML, XHTML, DOM

In English: Additional languages of the Web.
Briefly: JavaScript is a programming language that enables designers or

developers to build dynamic interactivity into their sites, further separat-
ing the Web from print. ECMAScript is a standardized version of JavaScript.
See Chapter 11, “The Joy of JavaScript,” for more particulars on this topic.
Cascading Style Sheets (CSS) is a standard that enables designers to con-

trol online layout and typography. Like HTML, its basics are extremely easy
to learn, though its subtleties elude many designers (as well as many
browsers). See Chapter 10, “Style Sheets for Designers.”
XML is a simplified version of SGML, designed for use on the Net. As of this
writing, it is most often used to deliver database-independent query results
between incompatible software applications. It is not yet universally sup-
ported in web browsers, though XML 1 is fully supported in Netscape 6, and
much of it is supported in IE5 and Opera 5. As a web designer, at least for
the next few years, you will hear about and see XML, but you will not be
called upon to create it—unless you begin marking up your web pages in
Extensible Hypertext Markup Language (XHTML).
08 0732 CH05 4/24/01 11:18 AM Page 128
XHTML is essentially an XML version of HTML that works in most browsers.

It is currently the W3C-recommended markup language for creating sites,
though most sites as of this writing are still created in HTML. The differ-
ences between HTML and XHTML, from the “writing the code” point of
view, are rather small, like Japan, though the implications of XHTML are
rather large, like China.
The Document Object Model (DOM) is a web standard that lets these other
standards “talk to each other” to perform actions. (For more about this, see
Chapter 11.)
With the increasing specialization of the Web, designers are no more

expected to master all these technologies than Rabbis are expected to fry
bacon. Web designers should learn CSS (which is easy), and most learn
enough JavaScript to be dangerous. Developers rather than designers will
likely do the XML and DOM programming as well as most of the heavy-duty
JavaScript/ECMAScript. The longer you work in the field, the more knowl-
edgeable you will become about these standards, but few employers will
expect you to have more than rudimentary awareness of most of this stuff.
Web page
As explained in Chapter 1, “Splash Screen,” a web page is a type of elec-
tronic document, just as a Microsoft Word file or a Photoshop document
is, except that a web page does not require any particular brand of soft-
ware for someone to open and/or use it. And that is the glory of it, broth-
ers and sisters. Developers and designers build web pages in HTML but, as
noted above, they also use stuff besides HTML, which we’ll talk about in
the relevant technical chapters.
Website
A website is a collection of related web pages published on the World Wide
Web.
In English: Click here.
08 0732 CH05 4/24/01 11:18 AM Page 129
Additional terminology
Tech terms will come up like last night’s chili burritos throughout your
career. You can always look up the latest buzzwords (or refresh your mem-
ory about what you have learned) by turning to the Computer Currents
High-Tech Dictionary at www.currents.net/resources/dictionary/.
ROLES AND RESPONSIBILITIES IN THE WEB

WORLD
As a web designer, you will be responsible for creating the look and feel of
websites—or portions thereof. Web designers may create menu bar icons
for sites designed by other designers on their team, or they may create ani-
mated ad banners for sites designed by others. Hey, you’ve read Chapters
2 and 3—you know the deal. (If you don’t, read the next chapter, which
describes the web designer at painful length.) Meanwhile, this seems as
good a place as any to familiarize yourself with some of the other players
on your team.
Web developer/programmer
The web developers on your team will be responsible for the technical
implementation of the site. You might hear them talk about Perl, Java, ASP,
PHP, SSI, XML, ColdFusion, and other technologies. Just smile and nod as if
you get it.
Most sites seamlessly fuse design, content, and interactivity. For that to
happen, teamwork is needed. You don’t have to understand how develop-
ers work their magic any more than developers need to possess design or
writing skills. But thoughtful collaboration and mutual respect for each
other’s disciplines are required to create functionally and aesthetically
superb sites.
Many developers have their roots in UNIX. Some are old hippies; others look
like preteen rejects from the cast of The Matrix. With the frantic need for
qualified personnel, developers also might come from the ranks of tradi-
08 0732 CH05 4/24/01 11:18 AM Page 130
130 WHO: The Obligatory Glossary: Roles and Responsibilities in the Web World
tional information technology (IT) services. Many of these people are won-
derful, but some have a strong bias toward particular technologies and
generally do not approach web development with a “Webby” mindset—by
which we mean a preference for open standards and accessibility.
For instance, IT-trained developers with roots in Microsoft-only shops

sometimes employ technologies that leave Mac, Linux, UNIX, and OS/2
users out in the cold. This is because they don’t know any better; they are
as trapped by their training as that sad little boy who shoots puppies. (We
now make good on our earlier apology to People for the Ethical Treatment
of Animals. You see, this book is really very skillfully woven together in spite
of its strange, dreamlike quality.)
Before accepting a job, be sure to check the company’s offerings using at

least one of these non-Windows operating systems. If the sites fail, the
developers may be biased in favor of proprietary technologies without real-
izing the harmful nature of that predisposition.
Designers who wish their work to reach the broadest possible audience
might want to think twice before accepting a job at a place like that. We’ve
even seen shops where Mac-using designers can’t send email due to Win-
dows-only gateway issues. This is not intended as a slam at Microsoft’s
many fine products, two of which were used in the creation of this book.
It is simply cautionary advice for the job seeker. In our opinion, because
closed systems lock out millions of potential users, serious web developers
prefer open standards.
Project manager
These team members are like technologically savvy account executives.
They help articulate the client’s needs, develop schedules (timelines) and
budgets, and are responsible for keeping the project on track. Just like the
account executives you might have worked with in your traditional design
career, project managers are usually good people with stressful jobs. As you
used to do with account executives, you must employ tact and patience to
negotiate with these folks.
Project managers will often produce things called Gantt Charts, which are
frankly little more than fancy work schedules. Say “thank you” when you
get these. It makes them feel good about themselves.
08 0732 CH05 4/24/01 11:18 AM Page 131
Do not actually look at the Gantt Charts, however. They will only frighten
you and make you feel hopeless and uncreative. Don’t worry about missing
any deadlines. The project managers will be “casually dropping by” your
cubicle every 15 minutes for the next 40 years, and you won’t miss a sin-
gle deadline. See, what did we tell you? They’re exactly like account
executives.
Systems administrator (sysadmin) and network

administrator (netadmin)
Sysadmins and netadmins are also called network engineers, database
administrators, directors of web development, webgods, UNIX guys, NT
guys, Linux guys, and geeks (formerly often called “webmasters”)—these
are the people who run the server (computer) that houses the site you’re
developing. They also run the staging server where you might build the site
before actually “publishing” it to the Web.
In some companies these folks are woefully underpaid juniors with exten-
sive computing knowledge. In others, they’re woefully overpaid juniors
without any computing experience at all, which is why many old-school
webmasters now call themselves systems administrators, network engi-
neers, database administrators, and so forth, leaving the webmaster
moniker to the temp who answers email about the site.
Most sysadmins are senior employees in charge of a staff. In small compa-

nies, they also might be the folks you go to when you need software
installed on your computer or if you’re having trouble with your email. In
larger companies, an IT person typically handles those responsibilities. In
some companies, sysadmins are also developers and in others, they are not.
Some companies call their sysadmins developers. Some call their juniors
seniors—because titles are easier to come up with than salary. Some say
love is like a flower. Don’t let any of this drive you mad. Above all, respect
the sysadmins. Without them, we’d have no Web.
Web technician
Web technicians, also called producers, web producers, HTML jockeys, web-
monkeys, web practitioners, HTML practitioners, design technicians, HTML
technicians, geeks, and many other things, are folks who do a job similar
to that of the studio people in an ad agency. As studio people take an art
08 0732 CH05 4/24/01 11:18 AM Page 132
132 WHO: The Obligatory Glossary: Roles and Responsibilities in the Web World
director’s comp and make technologically-oriented changes to it so that it

can be handed off to a printer, web technicians take a web designer’s Pho-
toshop comp, cut it apart, and render it in HTML, JavaScript, and other lan-
guages as needed. They also will render the graphic elements in
web-appropriate formats.
If you were wondering, the difference between web technicians and web
developers is largely a matter of experience, knowledge, and salary. A web
technician may cut your comp apart and write HTML; a web developer is
more like a technology designer who envisions powerful transactions and
writes advanced code in several different languages to bring those visions
to life.
Web developers are as critical to the process as lead designers, whereas

web technicians are more like junior designers. Junior designers might cre-
ate buttons or develop alternate color schemes under the supervision of a
senior designer; similarly, web technicians generally do lower-level pro-
gramming tasks than web developers. Some don’t program at all but sim-
ply use cut-and-paste JavaScripts (as do many web designers).
In smaller companies (or in large companies when there is a time crunch),

there are no web technicians; web designers execute their own designs in
HTML, JavaScript, and other languages as needed. This book will prepare
you to do that part of the job, making you that much more employable and
giving you that much more control over the process. (And as designers, we
all like control.)
Even if you always work in large companies, your knowledge of these

processes will enable you to work closely with web technicians, often in a
supervisory capacity—as if they were junior designers helping you execute
your campaign. It also will enable you to self-publish your creative work if
you find, as many of us do, that the Web is the greatest thing since sliced
bread, and you have an urge to do creative work even after hours.
We started this chapter by mentioning that titles are often confusing in

this business. The same thing holds for job responsibilities. While there are
plenty of junior and mid-level web technicians, there are also web devel-
opers who handle these tasks. Regardless of anyone’s stature, it goes with-
out saying that you should be respectful to all your teammates because
that makes life better.
08 0732 CH05 4/24/01 11:18 AM Page 133
And speaking of you…
YOUR ROLE IN THE WEB

Will be covered in the very next chapter. Go there.
08 0732 CH05 4/24/01 11:18 AM Page 134
09 0732 CH06 4/24/01 11:18 AM Page 135
chapter 6
What Is a Web Designer,

Anyway?
WE’VE EXPLORED THE STRENGTHS AND LIMITATIONS of the Web as a unique

medium; considered architecture and navigation as key components of the
user experience; glanced at the medium’s history; defined technical terms;
and examined the roles played by your coworkers. Maybe it’s time to look
at your job on the Web. We’ll start with a working definition.
Definition
Web designers are professionals who solve a client’s communication problems

and leverage the client’s brand identity in a web-specific way.
Complementing this focus on the client’s needs, web designers must think like
the site’s anticipated audience. They foresee what visitors will want to do
on the site and create navigational interfaces that facilitate those needs.
Pretty dry stuff, we’ll grant you, but like marital bliss, it’s better than it
sounds.
How does all this fancy talk break down in terms of daily tasks? Below is a
summary of deeds you’ll do during the web development project life cycle.
In Chapter 7, “Riding the Project Life Cycle,” we delve into details.
09 0732 CH06 4/24/01 11:18 AM Page 136
136 WHO: What Is a Web Designer, Anyway?
Through the project life cycle, the web designer will need to:
■ Understand and discuss the underlying technology—its possibilities

and limitations as well as related issues—with clients and team
members.
■ Translate client needs, content, and branding into structured web-

site concepts.
■ Translate projected visitor needs into structured website

concepts.
■ Translate website concepts into appropriate, technically
executable color comps.
■ Design navigation elements.
■ Establish the look and feel of web pages, including typography,

graphics, color, layout, and other factors.
■ Render design elements from Photoshop, Illustrator, and other visual

development environments into usable elements of a working
website.
■ Lay out web pages and sites using HTML and other web development
languages.
■ Organize and present content in a readable, well-designed way.
■ Effectively participate on a web development team.
■ Modify graphics and code as needed (for instance, when technolog-

ical incompatibilities arise or when clients’ business models change—
as they often do in this business).
■ Program HTML, JavaScript, and style sheets as needed. In larger

agencies, this work is often performed by web developers and tech-
nicians (see Chapter 5, “The Obligatory Glossary”), but the accom-
plished web designer must be ready to do any or all of these tasks as
needed.
■ Try not to curse browser makers, clients, or team members, as obsta-

cles are encountered throughout the process. (Well, go ahead and
curse browser makers if you want to.)
09 0732 CH06 4/24/01 11:18 AM Page 137
■ Update and maintain client sites as needed. Though this job, too,
often falls to web technicians or producers, don’t think you’re off the
hook. You’re never off the hook.
WHAT WE HAVE HERE IS AN OPPORTUNITY

TO COMMUNICATE
The work of web design involves understanding what your clients wish to
achieve, helping them refine their goals by focusing on things that can be
done (and are worth doing), and ultimately translating those goals into
working sites.
While interacting with clients, you’re also interfacing with research and
marketing folks to find out who is expected to visit the site and what they
will demand of it. You’ll be translating the anticipated needs of projected
visitors into functional and attractive sites—and hoping that visitors want
what your client wants them to want. (Try saying that with a mouthful of
peanut butter.)
If visitors seek in-depth content, but your client envisioned the site mainly
as a sales channel, either the client has fundamentally misunderstood his
market (it happens), or your design is sending the wrong messages. To build
sites that clearly convey what they are about and how they are to be used,
you must first communicate unambiguously with clients, marketers, and
researchers.
The site can’t communicate unless the people who build it communicate.
Ever try to design a logo for a client who could not articulate the target
market, product benefits, or desired brand attributes? The same problems
crop up in web design unless you are blessed with great clients or are will-
ing to work with the ones you have. Listening may be the most important
talent you possess. If your listening skills have grown rusty, you’ll have
plenty of meetings in which to polish them.
Good web designers are user advocates as well as client service providers.
They are facilitators as well as artists and technicians. Above all, they are
communicators, matching client offerings to user needs.
09 0732 CH06 4/24/01 11:18 AM Page 138
138 WHO: What Is a Web Designer, Anyway?: What We Have Here Is an Opportunity to Communicate
As designers, we often look down on clients for reacting according to their

personal taste (“I don’t like bold type”) instead of viewing the work through
the eyes of their intended market (“That’s just what our customers are look-
ing for”). But web designers commit the same offenses. Some of us become
so enamored of our aesthetic and technical skills that we end up talking to
ourselves or sending encoded visual messages to our fellow web designers.
As a design professional, you are presumably free of this affliction most of

the time. (If not, you’d have found some other line of work by now). Retain
that focus (Who am I talking to? What are they looking for?) as you pick
up the tools of your new trade. If you emphasize communication above all
other goals, you will find yourself enjoying a significant competitive advan-
tage. You’ll also design better sites.
Let’s expand our definition of the web designer’s role.
Definition (Revised)
A web designer is responsible for the look and feel of business-to-

business and business-to-consumer websites. Web designers solve their
clients’ communication problems, leveraging brand identity in a web-specific
manner (in other words, in a manner that respects the limitations and exploits
the strengths of the Web). A web designer understands the underlying tech-
nology and works with team members and clients to create sites that are visu-
ally and emotionally engaging, easy to navigate, compatible with visitors’
needs, and accessible to a wide variety of web browsers and other devices.
The Definition Defined

Let’s break this definition into its components:
A web designer is responsible for the look and feel of business-to-

business and business-to-consumer websites.
Look and feel

Just as in print advertising, editorial work, and graphic design work, the
look and feel reflects the client’s brand, the intended audience, and the
designer’s taste. Is the site intended for preteenage comic book fans? Is it
09 0732 CH06 4/24/01 11:18 AM Page 139
a music site for college students? An entertainment site? A corporate site?

An informational or shopping site for a wide, general audience? Is it
intended to reach an international visitorship? Or just people from Ohio?
(Is visitorship a word?)
As with any design assignment, you first find out all you can about the
client’s brand and the audience the client intends to reach and then make
appropriate decisions. The terrain will be familiar to you. It includes choos-
ing typefaces, designing logos, selecting or creating illustrations or photo-
graphs, developing a color palette, and so on. As we discussed in Chapter
2, “Designing for the Medium,” these familiar tasks change a bit when
applied to the Web because the medium embraces certain things (flat
color fields, text) while hiccuping on others (full-screen graphics, high-
resolution images and typography).
More significantly, “look and feel” decisions extend beyond traditional

graphic design and art direction to encompass site-wide navigational
architecture (as discussed in Chapter 3, “Where Am I? Navigation & Inter-
face”). Technological issues play their part as well. A site in which database
queries generate results in HTML tables will have a different look and feel
than a more traditional content site, or one created in Macromedia Flash.
The technological choice does not dictate the look and feel: It can be
any kind of HTML table-based layout, any kind of text layout, or any
kind of Flash-based design. The choice of technology merely establishes
parameters.
Business-to-business
Business-to-business means one company communicating with another or
selling to another. Annoying dot-com types and techno-journalists refer to
this as B2B.
The B2B category includes intranet sites (the private, company site of
Ogilvy & Mather or Pepsi Cola) and extranets (a steel company’s site linked
to a broker’s site linked to the sites of five customers). Flip back to Chap-
ter 5 if these terms make you edgy. Though this part of the web business
is hidden from most folks, it is vast and growing. There’s no doubt that in
your web career, you’ll be asked to design some B2B sites. You’ll also have
to avoid slapping people who say “B2B.”
09 0732 CH06 4/24/01 11:18 AM Page 140
In fact, we’d like to apologize right here for using acronyms such as B2B
and B2C. They annoy us as much as they do you. But you might as well get
used to them because you’ll be hearing them constantly at your job.
Besides, as annoying as these acronyms are, they’re not nearly as nerve-
wracking as ubiquitous venture capitalist phrases such as “burn rate,”
“built to flip,” or “ad-sponsored community play.”
We’ve never understood why these phrases arise, let alone how those who
talk that way manage to avoid being beaten with large polo mallets on
a daily basis. Our theory is that such phrases make the speakers feel impor-
tant. As you can probably tell, we didn’t have much to say about the
business-to-business category because, basically, web design is web design
regardless of the acronym attached to a particular category. Vanilla, choco-
late, or strawberry—ice cream is ice cream, Jack. (But do look back at Chap-
ters 2 and 5 for hints on coping with intranet-design-specific issues).
Business-to-consumer
When most folks think of the Web, they form a mental picture of business-
to-consumer sites such as Amazon.com—a business that sells products to
consumers like us. Not all B2C sites are overtly hawking products.
Yahoo.com is a B2C site. Yahoo! (the business) provides web users with
information. It isn’t selling anything per se, but it’s still B2C because it
speaks to consumers and is open to all. It’s not hidden on a private network
and password-protected, as a B2B site would be. The B2C segment is the
most visible part of the web. (We apologize for using the word “segment.”)
Solve Communication Problems

Let’s continue with the next part of the job description:
Web designers solve their clients’ communication problems, leveraging

brand identity in a web-specific manner (in other words, in a manner
that respects the limitations and exploits the strengths of the Web).
09 0732 CH06 4/24/01 11:18 AM Page 141
Using HTML to lay out web pages does not make you a web designer—nor
does making pretty pictures in Photoshop. A web designer, like any other
designer, is a communications professional who solves problems. Just as a
CD cover says something about the music it contains, the band that cre-
ated the music, and the likely customer, so the site must clearly commu-
nicate its structure, content, and purpose in a way appropriate to a specific
audience.
Gosh, haven’t we made this point before? Yes we have. And yet many web
designers will read these words, nod their heads sagely (or maybe just nod
off), and then continue to create sites whose appearance has nothing to
do with the product, user, or brand.
Brand identity
As a designer or art director, you know what this means. But what does it
mean on the Web? In simplistic terms, and on the most basic level, it means
the same kind of work you’ve done all your professional life: Make the logo
bigger. Use the client’s color palette.
But on a deeper level, the web designer doesn’t merely “use the client’s col-
ors” and slap the client’s logo on a web page. The web designer uses the
site to express and extend the client’s brand identity.
In Chapter 3 we discussed the way IBM’s brand positioning as a solutions

company influenced not only the site’s look and feel, but also the depth
and nature of its architecture and the type of enabling technology
employed in its construction (see Figure 6.1). Good web designers are
always thinking beyond the surface, extending and translating the brand
through function as well as form.
Web-specific
No surprises here. In the case of the IBM site, “leveraging brand identity in
a web-specific manner” means designing a site that provides solutions, not
problems. Clear navigation and a search engine that works help the site
support this aspect of the brand. This is an example of using the Web’s
strength as a searchable database to convey brand attributes.
09 0732 CH06 4/24/01 11:18 AM Page 142
Figure 6.1
Did the designers of IBM’s
website (www.ibm.com)
succeed in their quest to
translate the IBM brand
to the Web? Front-page
graphics tell only part of
the story. The site’s func-
tional performance tells
the rest. Web design
encompasses graphic
design but extends
beyond it.
Restrictions of the Medium

Every medium has limitations. This book, for instance, lacks hyperlinks and
a soundtrack. You can’t bookmark a motion picture (at least, not in the
theater—the management might complain), and you can’t save printed
magazine images to your desktop (though you can often save newsprint to
your fingertips).
The Web’s restrictions, as well as its strengths, were discussed in Chapter

2. Respecting those limitations and playing to those strengths is a key dif-
ference between design and web design. A web page that ignores the
medium’s restrictions (for instance, by forcing the viewer to download
100K of bloated imagery) or that fails to play to the medium’s strengths
(for instance, by offering limited interactivity), may be visually beautiful—
but it will still be poor web design.
Let’s look at the last part of our definition:
A web designer understands the underlying technology and works with

team members and clients to create sites that are visually and emo-
tionally engaging, easy to navigate, compatible with visitors’ needs, and
accessible to a wide variety of web browsers and other devices.
09 0732 CH06 4/24/01 11:18 AM Page 143
Technology
Web designers have a lot to say about the appropriate technological level
for sites they design. Choosing appropriate technology is part of your
job as brand steward and user advocate. Consider the following:
■ You wouldn’t design a general shopping site that depended on the

visitor having the Flash plug-in, the latest version of Internet
Explorer, or a particular operating system because you’d lose many
customers that way. The owners of Boo.com, a technologically over-
wrought shopping site, learned this the hard way when their busi-
ness imploded in 2000.
■ On the other hand, when designing a gaming site for Playstation or
an entertainment site for a high-tech sci-fi flick, using Flash
(or designing for newer, more capable browsers) could be entirely
appropriate.
■ You or a developer on your team might have fun coming up with a

nifty Dynamic HTML (DHTML) menu geared for Internet Explorer 5,
Netscape 6, and Opera 5—three recent browsers that to greater or
lesser degrees support the World Wide Web Consortium (W3C) stan-
dard Document Object Model (DOM). You would not create a menu
like that, however, for a women’s health care center because patients
and their families are not going to download a new browser when
seeking medical help or information.
Technology choices are essentially decisions about who the site is for. As a
communications professional, you should cultivate an informed opinion on
this matter. If you don’t decide these issues for yourself, somebody else will
decide for you, which can have potentially tragic results.
It’s also worth repeating that even if you decide the site is primarily for
bleeding-edge web enthusiasts, you will want to create alternative pages
that are accessible to anyone.
09 0732 CH06 4/24/01 11:18 AM Page 144
Works with team members

Although sites are often driven by a lead designer and technologist (or a
lead information designer), web design is nearly always a group effort.
Think of your team members as friends. In fact, think of them as family.
You’ll probably see more of them than you do your friends and family any-
way. Then again, as a designer, you may already be used to that.
Visually and emotionally engaging

Like we have to define this for you.
Like that ever stopped us.
Beyond functioning appropriately for its intended use and supporting the
brand, if your site lacks visual appeal or a coherent and engaging message,
all but the most dedicated users will pass it by in favor of a more fulfilling
experience elsewhere.
“Form follows function” does not mean “form doesn’t matter.” Form mat-
ters a heck of a lot. Given two functionally equivalent sites, only one of
which delights the eye, where would you choose to spend your time? Okay,
you’re a designer. But given the same two sites, where would your Aunt
Martha choose to spend her time? Okay, well, yes, we forgot about Aunt
Martha’s problem. Anyway, you get the idea.
Visually appropriate does not mean visually unengaging. Most of the

screenshots in Chapter 3 are of appropriately designed sites, very few of
which are lackluster or emotionally unappealing. We adopt kittens but run
from buzzards and rats because, well, to be honest, because buzzards and
rats are filthy, disgusting animals—but also because kittens are cuter than
buzzards and rats. We idolize babies and movie stars for much the same
reason.
You did not go into design to make the world duller or uglier. Anyone who
tells you a functional site has to be visually plain is suffering from an emo-
tional problem. Don’t make their problem yours. (But don’t give them
ammunition by designing a beautiful but hard-to-use site.)
09 0732 CH06 4/24/01 11:18 AM Page 145
Sites cannot be emotionally engaging if they don’t have a clear purpose

and a distinctive, brand-appropriate look and feel. It also helps a great deal
if they’re well written. Few commercial sites are. If you end up supervising
budgets for some of your projects, be sure to leave money for good writ-
ers and editors. Great cinematography can only go so far when the script
is bad.
Easy to navigate
Refer to Chapter 3.
Compatible with visitors’ needs

Refer to this chapter’s previous discussion of the three partners in any web-
site (the designer, the client, and the end-user) and to Chapter 3, which
covers scenario development as a means of getting inside the user’s head.
We get inside the user’s head (to the best of our abilities, anyway) to struc-
ture and design a site that meets that user’s needs. Aside from your Uncle
Marvin’s personal home page, no site appeals to just one user. We construct
multiple scenarios to forecast the needs of multiple users.
Accessible to a wide variety of web browsers and other

devices
We’ve already pointed out that the Web is accessed by a wide range of
browsers and that each of them has peculiarities, also referred to as incom-
patibilities. (Other words are also used, but we gave up swearing for Lent.)
Until all browsers support a core group of common standards, you will have
to learn the ins and outs of each distressingly different browser and con-
firm what you think you know by testing your completed designs on as
many browsers and platforms as possible. (We’ll discuss testing in the next
chapter.) In addition, your sites might need to work in nontraditional
browsers and Internet devices such as Palm Pilots and web phones.
09 0732 CH06 4/24/01 11:18 AM Page 146
146 WHO: What Is a Web Designer, Anyway?: Can You Handle It?
CAN YOU HANDLE IT?

By this point, the job of a web designer may appear too difficult. How is it
possible to reconcile the needs of the user with the demands of the client
and the heritage of the brand—not to mention coping with bandwidth lim-
itations, browser incompatibilities, and the unknowable behavior of each
individual visitor? Is it really possible to do this job well?
Obviously, we think so. Here are some not-so-obvious reasons why.
For one thing, web work is teamwork. Project managers, developers, web
technicians, writers, producers, and other designers on your team will help
you keep your eyes on the prize.
Moreover, as a design professional, you already possess most of the skills

and talents needed to design great sites, including:
■ The ability to research your client’s products and end-users, creating

work that promotes the former while speaking to the latter.
■ A deep understanding of branding and identity.
■ A comfortable familiarity with the processes of learning from and

presenting to clients and colleagues. You know how to sell and when
not to. You’ve learned how to listen.
■ Maintaining schedules and deadlines. You deliver on time.
■ A thorough knowledge of design principles.
■ Expertise with digital design tools, such as Adobe Photoshop and

Illustrator.
You can count on your teammates. You can count on yourself. And the
process itself also will help you meet the goals you, your clients, and part-
ners set for each project. Virtually every web agency employs methodolo-
gies and processes to guide you and your teammates from the initial
meeting to the launch (and beyond). By a strange coincidence, you’ll start
learning about that very subject as soon as you turn the page.
10 0732 CH07 4/24/01 11:19 AM Page 147
chapter 7
Riding the Project Life

Cycle
IN HOLLYWOOD, THE DIRECTOR IS KING. No matter how brilliant the work of the
actors, producers, screenwriter, cinematographer, composer, editor, set
designer, or other professionals, when the lights go down it is the director’s
vision that fills our eyes and forces us to respond.
On the Web, compelling sites begin and end with the vision of a lead
designer or a small, high-level design team. Other professionals certainly
play invaluable roles in defining and executing sites, however. Sites would
not work at all without the efforts of information architects, programmers,
producers, systems administrators, writers, and quality assurance teams—
to say nothing of focus groups, testing groups, marketers, and the occa-
sional consultant. And then there’s the client, who not only foots the bill,
but also contributes marketing and product information, existing artwork
and promotional materials, and his own ideas.
But sites that transcend mere adequacy depend on the consistent vision of
web designers. That means you.
Design at this level is broad and deep. It does not end with the creation
of graphic design elements. In fact, it does not even begin there. It starts
with the first meeting and continues straight through the launch. Under
ideal conditions, it goes on to include training and maintenance. For web
designers to stay actively involved in every step of the process, they must
thoroughly understand how the process works—hence this chapter.
10 0732 CH07 4/24/01 11:19 AM Page 148
148 WHO: Riding the Project Life Cycle: What Is the Life Cycle?
Make no mistake: If you skip any part of the process, you pay for it later—
with a site that falls short of your vision.
This chapter sketches life in the trenches of web development. It empha-

sizes the value of a methodology, outlines the life phases of most web
projects, and explains the kind of contributions you’ll be expected to make
in each phase of the process. Living this life is exciting, rewarding, and
sometimes quite stressful. Reading about it is dull as dirt. If you feel like
skipping this chapter, we’ll understand. It will be here when you need it.
For instance, just before you take your next job.
WHAT IS THE LIFE CYCLE?

Every project, from an ad campaign to the development of a new car, has
a life cycle. In most shops, web designers are expected to see a project
through from the initial discussion phase to completion and updating. In
some shops, this is not required; but in those places, you’ll want to partic-
ipate anyway.
If you’re not actively involved in the project from conception to “baby’s first
steps,” somebody else will be making critical decisions for you. That person
may not understand or care about consumer psychology, web usability, or
the importance of design. By understanding and involving yourself in the
entire project life cycle, you’ll be able to keep the focus on practicalities,
aesthetics, the client’s goals, and the needs of the site’s potential audience.
In your design career, you’ve undoubtedly toiled on projects that were mis-
directed long before you were brought into the loop. Designers can solve
many problems, but they cannot undo fatally misguided business decisions.
As an advocate for the end user and a spokesperson for the needs of your
team, you must be present from the beginning to the end.
Some web shops are designer-driven; others have roots in information

technology (IT). All good shops recognize the importance of involving the
design group early and often. Many web agencies formalize this role of the
design group by incorporating it into their methodology.
10 0732 CH07 4/24/01 11:19 AM Page 149
WHY HAVE A METHOD?

All websites, from e-commerce projects to abstract multimedia experi-
ences, contain elements of two types of activities:
■ Information systems, involving computers and software
■ Communication design, including advertising and marketing

communications
Because of the size and complexity of today’s sites, web development

often resembles information systems projects or enormous advertising and
marketing campaigns. It’s not as big a job as coordinating the cast and crew
of Gladiator, but it can come surprisingly close.
Though estimates vary, it’s agreed that the majority of information systems
projects fail. In case you missed that, we’ll say it again. Most information
systems projects fail. Why do they fail? It’s generally because there is too
much stuff to manage and keep track of, including the following:
■ Scope (the size of the project)
■ Budget
■ Resources
■ Timelines
■ Functionality (the stuff the site is supposed to do besides look pretty)
To help manage such complexity, companies have available to them

a resource that reduces the amount of unpredictability and surprise in a
project. It is called a methodology. Every good company has one; no two
are the same.
A methodology outlines steps required for a successful project, making sure

no steps are missed and none are undertaken at the wrong time. A method-
ology also organizes these steps into phases. Phases help team members
group activities, recognize progress, and notice red flags. A sound method-
ology provides documented, consistent, proven, repeatable processes. Pro-
jects that follow such methodologies work because they avoid reinventing
the wheel.
10 0732 CH07 4/24/01 11:19 AM Page 150
150 WHO: Riding the Project Life Cycle: Why Have a Method?
With a method in place, the team is freed from having to develop unique
support tools and processes for each new project. Without a method, the
team is driving off-road, blindfolded, without a map. They may reach their
destination safely, but it will be six months too late. They may end up in
Timbuktu, trying to convince the client they’re in Kansas.
The following story is true: Once upon a time, a web agency with no
methodology agreed to take on a large but fairly simple project. The client
delivered the copy 3 months late (they all deliver the copy late). The copy,
when delivered, was completely unusable. The agency had to pay a team
of freelance writers rush charges out of its own pocket because the client
had vetoed a writers budget. The client restructured the entire site as the
last graphic elements were being produced, invalidating all development
and graphic design work done up to that point and causing everyone to
work through Christmas to make up the difference.
Two weeks before launch, the client changed his logo and corporate
colors. A week later he changed his business model. The client faxed revised
(atrocious) copy from his vacation home, and it had to be manually retyped,
edited by those now-deliriously-happy freelancers, and then put into HTML
by freelance web technicians.
Just before launch, the client’s boss (the CEO) was brought in to bless the
work. Apparently, nobody had apprised him of the project plans. The CEO
hated everything. The client halted all work and, fearful of losing his job,
refused to send final payment. Attorneys were brought in. Agency staff was
laid off to pay the attorneys. Then the freelancers sued the agency for non-
payment. More staff was laid off. War was declared in Bosnia; Pinkerton
did not return—all because the agency failed to follow a methodology.
Successful web agencies often fall so in love with their methodology that
they broadcast it on their corporate sites. Whether they call it “our
method,” “our process,” or “Uncle Joe,” the discussion of corporate method-
ology is duller than fungus. So why do so many web agencies fill their sites
with such wearisome stuff? It’s because clients have been burned when
working with web agencies that seemingly had no methodology at all. The
trumpeting of methodology carries an implicit promise of performance.
(“We won’t be late or over budget. Look! We have a methodology!”)
10 0732 CH07 4/24/01 11:19 AM Page 151
Every married couple takes vows, but many later break them. Similarly, the
existence of a written methodology is no guarantee that the company that
wrote it will practice it. Nevertheless, good companies do have methods
that work for them, and you will want to master the methodology of your
web agency. If possible, you will want to improve it. Every project tests the
validity of the company’s methods, thus every project presents an oppor-
tunity to improve the company’s methods.
WE NEVER FORGET A PHASE

Like all human endeavors, web projects may be broken down into phases.
Each phase involves particular, predictable activities and results. We’re not
speaking of the mysterious spark of creative inspiration here; we’re talking
about process and workflow. Breaking a web project into phases allows
companies to predict and plan for activities, ensuring that no steps are
skipped. Reusing processes from one project to another also increases effi-
ciency while reducing heartache, phone rage, and legal expenses.
Phases are plans, and plans are never static. Over the life of any project,
activities move from one phase to another; activities may span several
phases; and lines of delineation between phases may blur. Still, it is possi-
ble to sketch a general outline of the web project life cycle, which is what
we’ve done here.
The five phases of site development are as follows:
1. Analysis
2. Design
3. Development
4. Testing
5. Deployment
10 0732 CH07 4/24/01 11:19 AM Page 152
152 WHO: Riding the Project Life Cycle: We Never Forget a Phase
Analysis (or “Talking to the Client”)

In this phase, you will meet with the client as often as necessary to fully
understand what the client wishes to achieve with the project, to deter-
mine the best ways of meeting those needs, and to sell those solutions
to the client. You’ll also continually interface with fellow team members to
make sure these solutions make sense and can be executed.
Even before sitting down to brainstorm with the team, you must help the
client articulate and clearly define the site’s goals. Is the site selling some-
thing? If so, what is being sold and to whom? Is the site intended to serve
as a portal—if so, a portal to what and for whom? How will this portal dif-
fer from its competitors? If the idea stinks, don’t be afraid to say so. (First,
of course, do enough homework to be certain that the idea really stinks and
be prepared to offer the client a better idea.)
These responsibilities are not the web designer’s alone. Project managers,
information architects, and marketing folks will be all over these meetings,
but the web designer plays an essential part.
Indeed, the web designer is often the only person in the room who even
thinks about the end user. The project manager is scheming ways to get the
project done on time. The programmer is itching to try out some new
technology or lazily conceiving ways to reuse code from the previous proj-
ect. The technology director is fretting about server farms. The junior
designer is nursing a hangover, and the client is lost in fantasies of market
domination.
The web designer must help the client articulate objectives, both broad and
narrow, to begin delineating the project’s scope. If this work is not done up
front, it will haunt the project (and the whole team) later on.
In these early meetings, the web designer should be prepared to discuss

possible site structuring options, technological baselines, and related
issues. Even if these ideas change later in the process—and they will—the
web designer must be comfortable articulating possible solutions “on the
fly.” This begins establishing a client comfort level, which will be essential
throughout the process. If the client does not trust the web designer in the
beginning of the cycle, the project will begin to self-destruct further down
the line.
10 0732 CH07 4/24/01 11:19 AM Page 153
To summarize what we’ve just said: It is essential that the web designer
possess the ability to understand a client’s marketing goals and to discuss
potential issues and solutions with regard to design, site architecture, and
technology.
To assuage your fears, the only part of this that is new (from your per-
spective as a professional designer) is that technological issues have been
added to the equation—much as ink, paper stocks, and such are part of the
traditional design equation. You will learn what you need to know in this
book and on the job.
The early phase

Earlier, we urged you to get involved at the very beginning of the process.
There is one phase in which you cannot participate. That is the client’s own
analysis phase. You will not meet with your clients until they’ve sat down
first to figure out their needs. Ninety-nine times out of ninety-nine, those
needs will change once you’re involved in the process.
How does the analysis phase operate? Just as in traditional design projects,
it typically begins at the highest levels of detail and works its way down.
In initial meetings, the focus is on broad strokes (such as, “We’re a com-
munity for young women.”). As the project progresses, lower-level deci-
sions emerge (such as, “Should we put buttons or text labels on tertiary
search results pages?”).
Though most of us are happiest alone in our cubicles, staring at our mon-
itors and though many of us would rather undergo gum surgery than face
another meeting, in many ways this phase is the most critical and creative
part of the job. The movement, over successive meetings, from the general
to the particular takes place on many levels and extends beyond issues of
graphic design and technology.
Many times, even the most sophisticated clients have only a rudimentary
and confused idea of what they wish to achieve. In their own realm, they
are kings. On the Web, they are lost little children. If your background
includes marketing experience and if you have made yourself knowledge-
able about the Web, you can guide your clients away from vague or even
nonsensical plans and toward worthwhile, achievable goals.
10 0732 CH07 4/24/01 11:19 AM Page 154
Take a simple project. Your client wants to sell videotapes online. He has
lined up a supplier and a fulfillment house, and after a full two hours of
online experience, he is convinced that his site will be “the Yahoo meets
AOL of online videotape e-tailing,” whatever that means. Because his
daughter, an art major, showed him the Monocrafts site (www.yugop.com),
a brilliant and beautiful work done entirely in Flash, he figures his site
should have “something like that” as well—oh, and a chat room. He read
about those in an airline magazine while flying between Seattle and New
York last year. He then describes the in-flight movie.
We wish we were making this stuff up, but it happens all the time. Not that
this client is necessarily an idiot—he may be brilliant in his accustomed
sphere of business. He may even read French literature and know fine
wines. It’s just that the Web is a mystery to him, and he’s not used to admit-
ting ignorance on any subject, even to himself. With tact and kindness, you
and your team will guide him toward a workable plan. Six months from
now, if you do your job well, he may have a fine site that includes movie
reviews by Roger Ebert, streaming video clips of selected films, and a thriv-
ing movie lovers’ discussion area. But it can happen only if you work with
him during the sometimes painful early analysis phase.
Defining requirements
Before all else, the web team must define two types of requirements:
■ Technical. These include anticipated performance, bandwidth,

security issues, and so on.
■ Business. These include needs and constraints (having to accommo-

date first-time web users), as well as overall marketing objectives.
These requirements are summarized in documents with impressive-sound-

ing names such as “Functional Spec,” “Requirements Document,” or the
ever popular “Use Cases Document.” And the fun doesn’t stop there: par-
ent documents beget baby documents—all of which will be used to guide
initial development, and none of which are carved in stone. The more stuff
you figure out, the more you realize you have yet to figure out. Digital proj-
ects kill more trees than the Daily News. You will be buried in paper. Read
it, absorb it, and set it aside.
10 0732 CH07 4/24/01 11:19 AM Page 155
Happy families are all alike, but every web project is different. Generally,
though, the purpose of early analysis is to define goals, determine con-
straints and requirements, and establish trust. Without goals, constraints,
and requirements, it will be impossible to know if the project is on target.
Without trust, you are looking at months of sheer Hell. With trust in place,
you may still be looking at months of sheer Hell, but you have a better shot
at enjoying the process and creating something useful.
If this sounds familiar, it should. The only difference between analysis in

traditional design and analysis in web design is the medium itself. Instead
of die-cuts or film transfers, you’ll be discussing bandwidth and browsers.
How it Works: Analysis in Action
Dishes Plus is a regional chain of successful retail outlets, known for its
reasonable prices, wide variety, and “break proof” guarantee. Dishes Plus has
decided to sell its product online. Naturally, you learn about the company’s
existing business, its competitors, and its brand image before the meeting. You
and your team help Dishes Plus define large goals (selling dishes), small goals
(branching out into soup tureens), and in-between goals (establishing a bridal
registry division).
You find out about the company’s audience (mostly women/mostly men,
young/old, urban/rural) and sketch the impact that may have on technologi-
cal and design considerations. If Dishes Plus has a loyal audience of people
over 50, tiny type is out, and plug-in based multimedia is probably out as well.
If selling is key, technological considerations leap to the forefront and should
be examined carefully.
How many clicks from expression of interest to final sale? If the inventory is
vast, a search engine will be needed. If Dishes Plus shoppers tend to spend
hours poring over the goods, artificial intelligence may be called into play on
searches (“If you like the Dixie Deluxe Classic Set, you’ll love the Colonel’s Tea
Service”).
Does Dishes Plus anticipate an overseas market? You might need to consider
building the site in several languages and using iconography to facilitate nav-
igation by non-English speakers. Do details matter? You can’t assume that the
client’s photography is up to snuff. You may need to budget for a good shooter,
conversion from photography to digital images, and a database to store and
serve the relevant images.
10 0732 CH07 4/24/01 11:19 AM Page 156
How does the database work? Your developers know. Meet with them sepa-
rately and then bring one or more to the next client meeting.
The possibilities are endless—when you first enter the room. After several suc-
cessful analysis meetings, the possibilities should have focused into a set of
meaningful and achievable goals. If you’re still talking in generalities after two
or three meetings, you’re not doing your job. If you’re talking in generalities
after four or five meetings, the client is not serious. Timelines with cash con-
sequences can sober up most clients. Have an attractive, friendly project man-
ager explain to the client the additional costs incurred as his indecisiveness
causes deadlines to shift.
Design
The design phase is a simple word for a heck of a lot of activity. The process
nearly always unfolds something like this:
■ Brainstorm and problem solve with your team.
■ Translate needs into solutions.
■ Sell ideas to the client.
■ Identify color comps to be developed.
■ Create color comps and proof of concept.
■ Present color comps and proof of concept.
■ Revise and repeat as necessary.
■ Receive design approval.
Brainstorm and problem solve

As soon as your team has a clear understanding of the client’s business
problems, goals, constraints, and requirements, you can begin brainstorm-
ing solutions on your own, in partnership with other web designers, and in
meetings with developers, producers, and information architects.
10 0732 CH07 4/24/01 11:19 AM Page 157
A project manager will join the team if she has not already done so. Make
her your best friend. While you conceive grand notions or are daydreaming
in Adobe Illustrator, she will be keeping track of and documenting sched-
ules, deadlines, goals, and progress. We wouldn’t last a minute in her job.
Nor would most “creative” folks. Respect her for doing what you would pay
not to do.
Methods of brainstorming vary. Some groups like to shout out ideas, writ-
ing everything down on a whiteboard. Others like to go off in small groups
and then reassemble to critique each other’s ideas. Sometimes you sit in
the corner and type out ideas. Sometimes you draw on a traditional sketch-
pad. Some agencies dictate how the process should work; others let you
figure it out for yourself.
Translate needs into solutions

The web designer and other team members will collectively come up with
a number of solutions, which will then be narrowed down by group con-
sensus, creative director fiat, or both. These solutions may be articulated
internally through any combination of rough design sketches, internal Pho-
toshop comps, written documents, or wireframes (functional visual story-
boards showing the proposed site elements in relation to each other, but
not in any way indicating how they will eventually look and feel).
To present these ideas to the client, you can once again use any of the
following:
■ Rough design sketches

■ User interface documents
■ Creative briefs
■ Pencil sketches
■ Wireframes
■ Color comps
10 0732 CH07 4/24/01 11:19 AM Page 158
Sell ideas to the client

Using any or all of the tools just listed, the team presents their projected
solutions to the client, answering questions, justifying decisions and meth-
ods, and discussing alternatives. As part of this discussion and “selling”
process, the designer should be able to:
■ Articulate technology limitations. Explain why the team supports a

particular solution and avoid committing to alternatives that won’t
work.
■ Articulate design considerations and decisions. As in a traditional

design project, explain the rationale behind various creative
decisions.
Articulating the limitations of technology and the needs of users can be

tricky. The web designer must be familiar with technological issues involved
in web development to be able to explain why the team supports a partic-
ular solution and to prevent impossible agreements and commitments.
Impossible agreements occur when the client asks for something that can-
not be done, cannot be done within the budget and time frame, or just plain
should not be done—and an inexperienced web designer or project man-
ager commits the company to fulfilling that unreasonable or impossible
expectation. Don’t laugh. Plenty of web design teams have met their doom
by committing to solutions that are technologically or graphically inap-
propriate, more costly than they’re worth, impossible to deliver within the
given time frame, or simply deeply stupid.
We were once asked to design the interface for a sophisticated, multi-user

business software program that ran in a web browser. Essentially, the prod-
uct was an intranet site with advanced functions rivaling that of expensive
proprietary software. Though a sales force had lined up dozens of large cor-
porate buyers, the developers were unable to deliver the product because
its scope kept shifting as the executive in charge came up with one “neat
idea” after another that he insisted on incorporating into the product. The
design team and sales force sat on their hands while the developers burned
out trying to fulfill constantly shifting objectives.
10 0732 CH07 4/24/01 11:19 AM Page 159
The executive then decided that the seemingly undeliverable product

would sell better if users could visually customize it to their liking. He asked
that a series of “skins” be developed, and the project manager added this
requirement to the mountain of unattained goals. The last we heard, the
product was still in development.
This kind of foolishness most often takes place in-house, where egos run
unchecked and projects can drag on forever without obvious financial con-
sequences—because those who do the work are on the company payroll
anyway. But it also can creep into traditional client-vendor relationships if
project managers accept impossible agreements.
That’s the worst-case scenario. The only-slightly-better scenario is that

your company will somehow fulfill the impossible agreement only to watch
the client fail because everyone shook hands over a really bad idea. The
client may want his e-commerce site visitors to enter personal data and
create a unique user account before even seeing what he has to offer. He
may request this at the last minute, and the web agency may manage to
fulfill the request on time and within the budget. But nobody will use the
site, and the client could bad-mouth the agency rather than admit his own
folly, thus harming your business for years to come.
Even if the client has only good things to say about you, you don’t want
your clients to fail, and you don’t want the press to associate your agency
with widescreen, Technicolor flops. It will take all your expertise at client
negotiation to avoid the Titanic effect (also known as the Boo.com effect).
But it’s better to face conflict than to knowingly deliver bad work.
The best-case scenario, of course, is to come up with and sell workable

solutions that offer real value to the audience your client wishes to reach.
How Not to Do It
“Because I know what I’m doing, and you’re a pathetic marketing flack who
wouldn’t know a good idea if it bit him on the thigh.”
10 0732 CH07 4/24/01 11:19 AM Page 160
How to Do It
“We considered that very solution, Burt, but these studies show it would take
50% more time than we have—and we found that companies who tried that
technique actually did worse than companies who did it the way we’re sug-
gesting. Marcie, could you show Burt that Tragic Failure Report you were shar-
ing with me before the meeting?”
A good web designer will sometimes lie in the service of a larger truth. Pre-
tend we didn’t say that, and we’ll pretend you don’t know exactly what
we’re talking about.
Selling your ideas is not limited to unselling bad ideas, of course. And it’s
also not limited to explaining technology. You’ll have the same design and
brand identity discussions you’ve been having for years. They still want the
logo bigger. They still prefer the obvious to the original. They still know just
enough to be dangerous to themselves and the project.
Identify color comps

You’ve finally determined the direction; in this phase, you figure out what
the client needs to see next. Typically, you’ll be creating comps of the web-
site’s front page and one or two internal pages. These comps are not func-
tional web pages; they are simply realistic renderings. At the same time,
you (or you and an information architect) will be developing storyboards
or wireframes outlining the flow of the site, from front page to order form,
from bulletin board to help page. You will not comp all these pages; you
simply need to know how they work.
Create color comps/proof of concept

Having identified the color comps necessary to prove the site concept, you
execute them in Photoshop or another design tool. Today’s web pages
almost always interact with the visitor—changing in response to mouse
movements and other events. Representing those interactive changes in a
comp may sound like a challenge, but it’s really not.
For instance, on most sites, an icon or menu item will change appearance
when the visitor’s cursor rolls over that icon. This change in appearance,
not surprisingly, is called a rollover. A comp can demonstrate the active
10 0732 CH07 4/24/01 11:19 AM Page 161
rollover state by showing one icon that is different from the others (“rolled
over”). To make the effect crystal clear, capture an image of a mouse
cursor and lay it on top of the “active” icon in Photoshop (Figure 7.1).
Figure 7.1
Is this a screenshot of an
active rollover on a web
page? Or is it a Photoshop
Present color comps and proof of concept comp? Only its hairdresser
knows for sure.
You have presented, articulated, and sold ideas to the client. Now you do
it again. The only difference is that the work is farther along in the process.
In addition to explaining the rationale behind design decisions and dis-
cussing the underlying technology, you also should be prepared to aurally
“sketch” what you have not yet comped up.
The client is interested not only in what you are showing; she is equally
interested in what you are not showing. “Are all the sub-pages like this
one? Will there be photographs on the message board pages, as there are
on the content pages? What happens if the search shows up empty? What
will that page look like? Does my hair look okay?”
You need to satisfy the client by describing (or “verbally storyboarding”)

these non-rendered pages. Prepare in advance. After all, you need to know
this stuff as badly as the client does. By having your answers ready, you’ll
shorten the approval process when it comes time to design the next stage.
(Client: “Oh, right, that’s the area we said would have the yellow menu bar.
Now I remember.”) You also will further instill client confidence in the
design team.
After the presentation, you will almost certainly need to make modifica-
tions. Web clients are no different than other design clients. They all have
needs they can’t quite articulate until they’ve seen some work. As in your
current job, you must know the difference between minor changes, which
may actually enhance the site, and major changes that could throw the
entire project off course. It is your responsibility to communicate the full
impact of suggested changes.
10 0732 CH07 4/24/01 11:19 AM Page 162
The presentation and revision process can go several rounds, demanding

your tact as well as your expertise. You must be able to respond positively
to client requests and return with a solution that demonstrates your
responsiveness, without jeopardizing the end product.
Receive design approval

The great day arrives! The client has signed off (well, except for one more
teeny tiny change). Now you gear up to begin translating your concept into
reality. This phase is known as development.
Development
In the development phase, web designers work with other team members
to translate site concepts into functional web pages. While you design
additional graphic elements, create Style Sheets, and possibly code web
page templates in HTML and JavaScript, producers will be marking up
dozens, hundreds, or thousands of pages, and developers will be working
to make the entire site far more functional than HTML alone allows.
Up until now, you’ve felt pretty certain about the way the site would shape
up. After all, the front page and selected sub-pages have been designed
and approved. Now, you must take the elements of all those pages and
apply them to every single page of the site. Sometimes you do all this work;
sometimes assistants or colleagues pitch in; and sometimes the work is
carried out by the equivalent of production artists, whose work you may
supervise. What is important in this phase is to maintain consistency.
Is the navigational menu on the right-hand side in all existing comps? Then
it should remain there as new pages are designed, unless there is an over-
whelmingly important reason to move it. (“It doesn’t fit on this page” is not
a legitimate rationale; it merely means you must work harder and rethink
that page. Sometimes it means you must rethink the entire site.) The con-
sistent location of navigational elements provides a vital pathfinder for vis-
itors. Imagine trying to find your way in a strange city where the street
signs kept changing color or location. No city would be that cruel or fool-
ish. Neither should any website. (Flip back to Chapter 3, “Where Am I? Nav-
igation & Interface,” for more on this subject.)
10 0732 CH07 4/24/01 11:19 AM Page 163
Client-branding elements also must be treated with consistency

from page to page. There are technological reasons for this, as well as psy-
chological ones.
Psychologically, if the logo is always 32 x 32 pixels and always at the top

left, visitors expect to see it there on all pages. Such consistency reassures
visitors that they are still in the same “place” on the Web. After all, the Web
is a fluid and limitless medium, and the client’s site is just a drop in that
vast ocean. Consistent branding orients web users; inconsistent branding
disorients them. Love your audience and provide the markers they need to
know where they are (and your clients will think you’re doing it for them).
Technologically, once a graphic element is cached in the visitor’s browser,

it need not be downloaded again. Because most visitors use slow dialup
modems, the less downloading, the faster the site and the better the user
experience. Thus, if the same 32 x 32 image appears on every page, there
is no need for additional downloads, and each page of the site will appear
that much faster. (Refer to the discussion in Chapter 2, “Designing for the
Medium,” regarding bandwidth and caching if you’ve forgotten how this
works or why it matters.)
During the development phase, you’ll do things such as:
■ Create all color comps
■ Communicate functionality
■ Work with templates
We provide tips and pointers in the following sections.
Create all color comps

As you have seen, the design phase demanded the creation of selected color
comps. During development, one or more web designers will create color
comps for all pages. Depending on client expectations, the design team also
may show these comps to the client for approval.
10 0732 CH07 4/24/01 11:19 AM Page 164
As the team creates each color comp, technicians or junior designers will
cut it apart in Adobe ImageReady or Photoshop 6, Macromedia Fireworks,
or another similar software program. This process converts the color comp
into component elements, and these are finally assembled into a working
web page built with HTML and other web languages or with a What You
See Is What You Get (WYSIWYG) web layout program, such as
Dreamweaver. This is not the only way to create web pages, nor (in our
opinion) is it necessarily the best way. It is the primary technique used in
most shops, however, and every web designer should master the process.
Communicate functionality
Refer to the previous discussion on rollovers or image swapping, as it
can be called. In some web firms, the web designer will code those image
swaps in JavaScript. In other firms, the web designer merely articulates a
desired effect (complete with Photoshop comps), and a developer or web
technician writes the necessary code.
Functionality can include CGI and Java (for forms, e-commerce, message
boards, and chat functions), JavaScript (for special interactive visual
effects as well as less glamorous browser detection, plug-in detection,
forms validation, and so on), plug-in technologies (Flash, QuickTime, or
RealVideo), and beyond.
The communication travels both ways. At times the technologist will

explain intentions or limitations to the web designer; at other times the
web designer calls the shots. Web designers are not expected to know Java
programming or MySQL. Many web designers do not even work in Flash.
What’s expected is that you know enough about these formats and lan-
guages to work with those who specialize in them, articulating your vision
or responding to the direction of others.
By the way, we despise the word “functionality” even more than we hate
phrases such as B2B or B2C. Alas, it seems to be the best word for the job.
Former English majors, check your emotions at the door. This business has
more buzzwords than a venture capitalist convention.
10 0732 CH07 4/24/01 11:19 AM Page 165
Work with templates

In some cases, sites change very little over time. More commonly, the site
you design functions as a placeholder or shell for ever-changing content.
Sometimes this changing content is managed by the web agency. If the
client updates infrequently, you can simply write new HTML and create new
images to accommodate occasional changes like these. More often, your
clients will update their own sites, with mixed results. (We’ll be whining
about that later in the book.)
Today, content is often changed dynamically, by means of various backend

technologies. In such cases, you are not so much designing pages as you
are templates—visual and markup placeholders in which content will be
updated by means of a publishing system or in response to dynamic data-
base technology. The work is essentially the same as “traditional”
web design but involves special considerations that will be articulated by
the technologists on the team. (See Chapter 12 for more.)
Design for easy maintenance

In the best of all possible worlds, the web design team retains control over
the site as it evolves over the months and years. Control is usually accom-
panied by a retainer fee, which is negotiated at the very beginning of the
process. In reality, more and more clients assume control over the site when
it is delivered.
Designers and coders should always create highly structured and well-
documented work, so that they can easily go back to it and update it with-
out hunting for missing files, debugging errant file references, and so on
(or so that their clients, upon assuming control of the site, can perform
these tasks without damaging the site).
Upon finishing the site, you’ll accompany it with a style guide and docu-
mentation. These will be much easier to create if the site’s file structure
and naming conventions make sense to begin with.
10 0732 CH07 4/24/01 11:19 AM Page 166
For you, this means titling the logo image “logo.gif” instead of “uglyswirl-
withstupidbevel.gif,” calling the November header graphic “nov_head.gif,”
and so on. For the web technician (or you, if you write your own HTML),
this means naming the disclaimer page “disclaimer.html,” storing images
in a single directory labeled “Images,” and so on.
If care is taken throughout the development process, then updating and

maintaining the site will be easy and logical, whether updates are per-
formed by you, a production person, or your client.
Testing
Though a web development team will test its product throughout the proj-
ect life cycle, many web projects plan for a distinct testing phase. In this
phase, the development team has the opportunity to test the deliverable
against the design and functionality specifications.
In some cases, real users may test a site. In other cases, a specially trained
testing team will do the job. Testing by real users usually tells you more
about the site. We often get the most useful feedback by showing work to
the guys who deliver sandwiches. Be elitist in choosing typefaces but dem-
ocratic in designing interfaces. The Web is for people, not for experts.
Regardless of the testing technique involved, team members must work

together to track down the source of problems and implement solutions.
We guarantee that there will be problems. For one thing, no two web
browsers interpret code and markup exactly the same way (see Chapter 2).
For another, what seems clear to you may be baffling to the people who
use your site. Web designers tend to live two years ahead of the curve; web
users, who actually have lives, tend to live behind the curve. You know that
little rotating box takes visitors back to the home page; visitors may not.
Testing will reveal problems in browser compatibility and user acceptance;
then it’s up to you and your team to solve those problems.
Deployment
You’ve completed the site. The client has signed off on it. The files have
been transferred to the web server. Think you’re finished? Not quite yet.
Successful projects demand a smooth transition from the web team to the
client.
10 0732 CH07 4/24/01 11:19 AM Page 167
The updating game

In the early days, clients viewed the web designer as a species of magician.
They knew they had to have a web presence, whatever that meant, and they
felt that you held their fate in your hands. Not only were they eager to
approve what you created (because it was all magic to them), they also
were more than willing to retain you as the perpetual updater and refresher
of their online identity.
Then came FrontPage, GoLive, and Dreamweaver—tools that theoretically

let “anybody make a website,” whether they knew what they were doing
or not. Now, the possession of FrontPage does not turn your client into
a web designer any more than ownership of a Roland Drum
Machine turns the neighbor’s kid into Keith Moon. But the ability to gen-
erate HTML, the language with which web pages are created, has con-
vinced too many clients that they can save a buck or two by
purchasing one of these web-editing tools and updating their sites them-
selves. The results are often disastrous, for reasons that will be obvious to
any creative professional, but incomprehensible to many clients, whose
esthetic sensibilities have been shaped by cooking up pie charts in Power-
Point. Not that we’re bitter.
There are several ways to manage the transition. In one of the better
scenarios, you’ve designed a database-generated site for a large client with
much money and created a custom publishing tool enabling them to add
fresh content to the mix without befouling the site.
Alternatively, instead of providing clients with a custom publishing tool,

you might hook them up with an existing product, such as Zope
(www.zope.org) or Allaire Spectra (www.allaire.com). Some of these tools
use standard web languages such as HTML and XML; many use custom
markup and are part of larger proprietary product families. Some are sim-
ple enough for a client to use; others require considerable developer
involvement, which is one way of keeping your finger in the pie—if that’s
your idea of fun. What you gain in billings you may lose in IT people, who
quit from the frustration of continually guiding clients through complex
processes requiring specialized knowledge. This, however, is your boss’s
worry, not yours (unless you own the web agency).
10 0732 CH07 4/24/01 11:19 AM Page 168
In still other cases, your client will assign an in-house person or team to
take over the site. Sometimes these folks are in-house designers with solid
web experience. Sometimes they’re overworked marketers with a copy of
FrontPage.
Regardless of how the site is updated and by whom, in this final phase of
development, you get one more opportunity to preserve your work and
serve your client by creating documentation, providing client training, and
maintaining contact on a consulting basis.
Create and provide documentation and style guides

“Care and Feeding” instructions accompany everything from puppies
to houseplants, and websites demand the same loving attention. It is
important to provide the client with detailed notes on the location of files,
the fonts and color palettes used, photographers or illustrators involved,
and so on.
As more and more clients plunge into the business of updating their own
sites, it is vital to provide them with every possible scrap of information. If
you don’t take pains with this postpartum part of the process, your client
may paint a moustache on your Mona Lisa or send visitors running for their
lives when a Style Sheet or JavaScript file is accidentally deleted.
Remember: A book design is a book design, a finished ad is a finished ad,

but a website is never finished, and the client can always louse it up. Do
everything in your power to save your clients from themselves.
By the way, such documentation should be created even if the web agency
retains control of the project (including updating and maintenance). After
all, six months from now, do you really want to scratch your head trying to
remember which font you used, where your navigational menu graphics
were stored, or which script was responsible for a given function? Of course
not. This documentation will be easier to create, and the site will be easier
to update if you’ve followed the advice given earlier in this chapter and
designed for easy maintenance by establishing and following logical nam-
ing and structural conventions.
10 0732 CH07 4/24/01 11:19 AM Page 169
Provide client training

Sometimes it is enough to tell your clients which fonts and colors you used.
Sometimes it is enough to tell your children not to play with matches.
Usually, it is not enough. That’s why, whenever possible, the designer and
other team members should have after-meetings to discuss the site in
detail and provide as much client training as possible.
Besides helping the client avoid ruining a beautiful site, in-house training
also sends the message that your company cares. Clients who know you
care will come back with additional projects and will tell their friends on
the golf course about the integrity of your company.
If your clients are going to be writing HTML or (bless us) creating new
images, it is worth sitting down with them, at their computer or yours, and
pointing out the fine nuances of what you’ve done. You might even buy
fonts for them (matching the fonts you used), install the fonts on the
client’s computer, and show them how to work with Extensis Suitcase or
Adobe Type Manager Deluxe.
You may feel ludicrous doing this, especially if the client is not a graphic
designer, but it’s foolish to underestimate other people’s creative potential.
Besides, if they’re going to do the work anyway, you owe them and your-
self every possible assistance.
This whole thing is fairly unsavory, we’re afraid. It’s rather like a dentist
training patients to extract their own teeth, but it is an aspect of the busi-
ness, and coping is better than lamenting.
Learn about your client’s methods

Training is often bi-directional. While explaining your methods to an in-
house peer (or turning a client into a junior web designer of sorts), you also
should learn as much as you can about the way your client will work with
the site. If possible, you should learn about the software your client will be
using. It’s highly unlikely that your client will create HTML and other web
markup by hand. Fortunately, the number of WYSIWYG web editors con-
sidered good enough to use is fairly limited, so you can learn the basics and
pitfalls of your client’s software of choice even if you never touch the stuff
yourself.
10 0732 CH07 4/24/01 11:19 AM Page 170
170 WHO: Riding the Project Life Cycle: Work the Process
We recently ran into a puzzling problem where the web typography we had
established via Style Sheets kept disappearing from the client’s site after
he took it over. We had written a Global Style Sheet, placed it in a secure
location, and instructed the client never to touch it. Yet every time he
updated the front page, the Style Sheet reverted to an early, inferior vision,
and the client was constantly contacting us to ask why the site was going
to Hell.
Eventually we discovered that a site maintenance feature built into the

client’s software was the culprit behind the Case of the Changing Style
Sheet. When the client updated his index page, his software program asked
if he wanted to “upload related files.” Because that sounded like a pretty
good idea, the client always clicked Yes. The program then automatically
uploaded dozens of files from his hard drive to the server. An old Style Sheet
on his hard drive was automatically replacing the newer one we had cre-
ated. We re-sent him the updated Style Sheet, instructed him to turn off
the site maintenance feature, and from then on, all was well.
WORK THE PROCESS

The process you’ve just read about varies by agency, but the general out-
lines and the lessons involved should hold true for most companies and
projects. Some agencies keep themselves fairly aloof from their clients and
manage to do wonderful work in spite (or because) of it. Others become
deeply involved with their clients, establishing long-lasting, trust-based
relationships.
Some hold their clients to ironclad contracts and schedules, while others
are loose and almost playful in their approach. Some shops show the client
exactly one comp—take it or leave it. Others cover the walls. Some agen-
cies charge astronomical fees merely to write a proposal; others write pro-
posals, design comps, and create storyboards on spec—a terribly ill-advised
approach, but not as rare as it ought to be.
10 0732 CH07 4/24/01 11:19 AM Page 171
The main thing to remember is that every phase, every step of the process,
is potentially empowering. If you use initial meetings to establish trust and
help sharpen the client’s vision, you will find yourself working on sites
worth designing—for clients who respect you instead of mistrusting and
fighting with you. If you use the design phase to fully explore possibilities,
you will come up with richer designs and avoid structural problems in the
implementation phase. If you cooperate with team members and your
client during the production phase, you will encounter fewer problems dur-
ing testing. If you train your clients respectfully, your best efforts will be
preserved, you’ll be able to look at your old sites without experiencing nau-
sea, and the credibility of your work will win you new and better projects.
10 0732 CH07 4/24/01 11:19 AM Page 172
11 0732 Part III 4/24/01 11:20 AM Page 173
Part III
HOW: Talent Applied

(Tools & Techniques)
8 HTML, the Building Blocks of Life Itself 175
9 Visual Tools 209
10 Style Sheets for Designers 253
11 The Joy of JavaScript 285
12 Beyond Text/Pictures 327
13 Never Can Say Goodbye 387
11 0732 Part III 4/24/01 11:20 AM Page 174
12 0732 CH08 4/24/01 1:22 PM Page 175
chapter 8
HTML, the Building

Blocks of Life Itself
AS WE’VE SAID THROUGHOUT THIS BOOK, HTML is a simple language for creating
documents that adhere to structured outlines.
<h1>Headline</h1>
<h2>Subhead</h2>
<p>Paragraph.</p>
<p>Second paragraph.</p>
<p>Third paragraph.</p>
<h3>Subordinate subhead</h3>
<p>Paragraph.</p>
<p>Second paragraph.</p>
<p>Third paragraph.</p>
<address>Contact information, copyright, date of publication</address>
Rocket science it’s not, nor was it intended to be. All great ideas should be
this simple. Notice that the tags (that’s what the lines are called—tags)
suggest their functions: <p> for paragraph, <h1> for first-level headline,
<address> for contact information. Notice also the fine symmetry in this
simple example. You open a <p> and you close it </p> when you’re done.
You open a <h3> subhead and </h3> close it before moving on to another
tag. In this way, the browser knows that one tag has closed before another
begins. In HTML, the closing of some tags is mandatory, while with other
12 0732 CH08 4/24/01 1:22 PM Page 176
176 HOW: HTML, the Building Blocks of Life Itself: Code Wars
tags it’s optional. That inconsistency has led to sloppy markup, which in
turn has caused browser problems, especially when other web standards
(such as CSS) begin to interact with your HTML. So it’s a good idea to close
most HTML tags whether it’s strictly required or not. (In XHTML, the suc-
cessor to HTML, all tags must be closed.)
CODE WARS
As five minutes of web browsing will show you, HTML has been twisted
every which way to enable web designers to create documents that are not
so logical in their construction nor so restricted in their presentation. Cas-
cading Style Sheets (CSS) and JavaScript are additional technologies that
enable designers and developers to create attractive, accessible, dynamic
web documents.
In theory, web designers should let HTML be HTML, using it merely as a

structured container for content, while relying on CSS to format pages. In
practice, web designers had to design pages long before CSS was invented,
so most of us developed methods for using HTML as a design tool. Even
after CSS was invented (1996), the first reliably CSS-capable browser did
not hit the market until 2000. As of this writing, support for CSS is still
tragically far from complete in many popular browsers. More about that—
including solutions—in Chapter 10, “Style Sheets for Designers.”
Table Talk
As a result of the Thousand Year March toward CSS compliance and while
waiting for better browsers, designers still use HTML for tasks it performs
reliably, if grudgingly, such as creating visual layouts by manipulating
HTML tables:


<table border=”0” cellpadding=”0” cellspacing=”0” align=”center”>
<tr>
<td>
<a href=”reading.html”><img src=”reading.gif” width=”20” height=”20” border=”0”
alt=”Reading”></a>
</td>
12 0732 CH08 4/24/01 1:22 PM Page 177
<td>
<a href=”writing.html”><img src=”writing.gif” width=”20” height=”20” border=”0”
alt=”Writing”></a>
</td>
<td>
<a href=”arithmetic.html”><img src=”arithmetic.gif” width=”20” height=”20” border=”0”
alt=”Arithmetic”></a>
</td>
</tr>
</table>

The previous code, in conjunction with the appropriate <img> images

(reading.gif, writing.gif, arithmetic.gif), will result in a clickable naviga-
tional menu in visual web browsers such as IE, Netscape Navigator, Opera,
and iCab. The table is used not to present tabular data (such as the con-
tents of a spreadsheet) but rather to hold images in place. Setting the bor-
der to “0” disguises the tabular structure to facilitate pure visual purposes.
Typing the width and height for each image helps the browser more quickly
calculate how the data is supposed to lay out on the page. The <ALT>
attribute in each image tag makes the content accessible to users of audio
browsers and nontraditional browsers such as Palm Pilots, as well as for
those who surf with images turned off.
XHTML Marks the Spot

We keep emphasizing that HTML is logical and orderly. Let us return for a
moment to the question of closing tags after they are opened. This prac-
tice may seem redundant, but there is a logic to it. Refer again to the pre-
ceding example. Say that your paragraph is followed by an image. If you
don’t close your paragraph </p> before starting the image <IMG> tag, the
browser has to guess whether you intend the image to be part of the para-
graph or to follow it. Depending on what the browser guesses, your image
might be preceded by a carriage return, or it might not be. If you’re using
a style sheet that includes leading (line-height in CSS parlance), the
browser might attempt to impose that leading on your image—or it may
not. These are merely the visual complications that can arise from some-
thing as simple as an unclosed tag. The structural ramifications can be
more serious.
12 0732 CH08 4/24/01 1:22 PM Page 178
178 HOW: HTML, the Building Blocks of Life Itself: Code Wars
HTML recommends that you close most tags, but it does not force you to
do so, and it does not uniformly recommend tag closure. Images <IMG>,
line breaks <br>, and list items <li>, for instance, are never closed in HTML.
Most older browsers will accept and attempt to display all kinds of shoddy
markup, including markup with unclosed tags that should be closed. And
some browsers choke on optionally unclosed tags when, according to the
rules of HTML, they should simply process the markup without qualms. For
instance, Netscape’s 4.0 browser refuses to display web pages with
unclosed table rows, even though the rules of HTML 4 state that table rows
need not be closed.
This inconsistency in HTML (and browsers) has resulted in sloppy markup

on a surprising number of professional sites. Living in filth promotes dis-
eases in human beings; slinging dirty markup can have equally dire effects
on the health of the Web, particularly as the medium attempts to move
forward.
For this reason, among others, in 2000 the W3C stopped evolving HTML
and came up with a new standard called XHTML. Don’t sweat it. For a web
designer’s purposes, XHTML is essentially HTML that forces you to close
your tags—including those (such as <img>, <br>, and <li>) that never
required such closure before.
We don’t want to confuse you with yet another acronym, but the reason
XHTML works this way is because XHTML only looks like HTML. It is actu-
ally the offspring of XML, which is the standard toward which the Web is
evolving. (Technically, XML is a meta-language. That is, it is a set of rules
for creating languages. HTML is a markup language based on the SGML
meta-language. XHTML is a reformulation of HTML using XML as the meta-
language. You don’t really need to know this, but it’s great at parties, par-
ticularly when you’re trying to make someone stop talking to you.)
Minding Your <p>’s and q’s

Instead of trying to grasp the mind-numbing sentences above, you can
think of XML as a much smarter—and necessarily more complex—
adaptation of the idea behind HTML; as Homo sapiens to HTML’s Cro-
Magnon, if you will; as a structured meta-language for containing data, if
you must. In XML, you construct your own tags. Not only that, an XML tag
12 0732 CH08 4/24/01 1:22 PM Page 179
“knows” what other tags on the page are doing. An HTML tag does not. Nei-
ther does an XHTML tag (at least, not yet). But XHTML follows the rules of
XML, chief among them the demand that all code be “well-formed.” Tags
that close after they open are demonstrating well-formedness in XML parl-
ance.
Exploring XML is beyond the scope of this book and is also beyond your
immediate job requirements as a web designer. In all probability you will
spend the next few years working with HTML or XHTML. In practical terms,
working with XHTML is just like working with HTML, except that it enforces
cleaner coding practices. HTML doesn’t care if you clean up your room.
XHTML does.
LOOKING AHEAD
What you need to understand:
■ Web designers must learn HTML, even if most of the HTML work at
their jobs is performed by web technicians. Web designers who
choose not to learn HTML will limit their creative thinking as well as
their employability. Frankly, few good web firms would hire a
designer who lacks at least basic knowledge of the technology that
drives the medium, though many would hire a great designer whose
initial HTML skills are merely adequate. Fortunately, HTML is very
simple and thus very easy to learn.
■ HTML was not created as a design tool, and within the next year or
two, we will no longer be using it as one. However…
■ …the Web is in transition, from an anything-goes “tag soup” to a
more usable and logical division of labor between technologies that
structure content (HTML, XHTML, XML) and companion technologies
that format its display (mainly CSS).
When you begin working at a web shop, you or your coworkers will likely
be formatting your pages in HTML (or XHTML) to make them compatible
with late 1990s browsers. You will use CSS as well, but initially only in lim-
ited ways (more about that in Chapter 10). Soon, CSS will do more and
more of the design work, and HTML and its successors will be used as the
Web’s creators originally intended.
12 0732 CH08 4/24/01 1:22 PM Page 180
180 HOW: HTML, the Building Blocks of Life Itself: Looking Ahead
Directory Assistance
On the Web, files are referenced by Uniform Resource Locators (URLs) like
this one:
http://www.populi.com/index.html
These URLs are to cyberspace what street addresses are to the real world—no
two addresses are exactly the same.
UNIX was used to serve most sites at the dawning of the Web, and conse-
quently, URLs follow the conventions of the UNIX Operating System (for
instance, URLs are case-sensitive because UNIX is case-sensitive). UNIX is
still used to serve heaps of sites due to its stability and its deep roots in the
history of the Internet.
Other popular web-serving operating systems include Windows NT and,

increasingly, Linux. Linux is a free, open-source version of UNIX. Macs also
can serve websites, though most companies prefer to host on UNIX or NT
because these platforms were designed for the job. That may change some-
what now that Apple has unveiled Mac OSX—its next-generation operating
system with UNIX underpinnings.
In UNIX, slashes separate directory names from each other and from docu-
ment names, and all web servers follow these conventions.
The names of web documents (including images, movies, and audio files)
generally end in a three- or four- character abbreviation that clues the web
server (and the browser) as to what they are (and thus, how they should be
handled). HTML documents end in .html (or sometimes .htm); JPEG images
end in .jpg (portrait.jpg); Flash files end in .swf (grandioseintro.swf); and
so on.
All web-serving platforms follow these conventions in naming. Windows 3.1

is limited to three-letter abbreviations, so .htm is used on that platform. Few
sites are served from Windows 3.1, however, and you are astronomically
unlikely to encounter Windows 3.1 servers in your professional career—at
least in the United States, Canada, and Europe.
Apache is a powerful web-serving platform with many conventions that

designers follow even if they don’t know why they’re doing so. For instance,
at the root level (www.populi.com), if you include a document titled
index.html, that document will open automatically when the visitor types
www.populi.com. This is why the ads can say “Visit www.spamulator.com”
instead of “Visit www.spamulator.com/index.html.” This is also why, in writ-
ing code, you can save bandwidth with URL references such as:
<a href=”http://www.nostril.com/”>
12 0732 CH08 4/24/01 1:22 PM Page 181
Instead of:
<a href=”http://www.nostril.com/index.html”>
It’s also why you can reference internal files in this way:
Visit the <a href=”contact/”>CONTACT</a> page.
Instead of:
Visit the <a href=”contact/index.html”>CONTACT</a> page.
Or:
Visit the <a href=”http://www.nostril.com/contact/index.html”>CONTACT</a> page.
The systems administrator can override this default if she desires, allowing
welcome.html (for instance) to serve as the default opening document. In
fact, welcome.html was the default opening document on many systems
before index.html gained ascendancy. The default page at www.zeldman.com
is still welcome.html. (The old CERN server used Welcome.htm, complete with
initial caps.)
These conventions vary by system. Internet Information Server (IIS), on

Windows, uses default.html or default.asp. Again, the systems administrator
is free to override any such default. (Pickledherring.html could be set up as
the default document if desired.) If your server or systems administrator
prefers a particular filename, you’ll be told about it on the job.
GETTING STARTED
There are plenty of books about HTML, and heaps of free online resources.
After all, what better place than the Web itself to learn about the markup
language with which the Web is created?
The beginner’s tutorials at Project Cool’s Gettingstarted.net

(www.gettingstarted.net) and Jay Boersma’s Web Work (www.ECNet.Net/
users/gallery/webwork/www.html) can teach anyone, even your Uncle Phil,
how to apply basic HTML tags to create simple web pages. A more detailed
tutorial, Ian S. Graham’s “Introduction to HTML,” may be found at
www.utoronto.ca/webdocs/HTMLdocs/NewHTML/htmlindex.html, and our
own “Ask Doctor Web” (www.zeldman.com/askdrweb/), online since 1995,
12 0732 CH08 4/24/01 1:22 PM Page 182
182 HOW: HTML, the Building Blocks of Life Itself: Getting Started
provides a readable overview on HTML and related technologies along with

the psychology of web use and similar topics. After you’re further along,
you are sure to enjoy Lance Arthur’s Design-o-Rama at www.glassdog.com/
design-o-rama/, a wittily written treatise that includes a good introduc-
tion to frames and JavaScript.
These are but five of many such resources online. Most of these resources
are noncommercial in nature. They exist only to share knowledge. We told
you the Web was different, didn’t we? (On the Web, this would be a hyper-
link back to Chapter 2, “Designing for the Medium.”)
Being noncommercial, such resources might not always be completely up-

to-date. For instance, parts of Ask Dr Web show their age visually, and in
places, the advice offered might not be up to current standards. Neverthe-
less, they are all excellent places for those who don’t know their HTML from
their elbow to begin absorbing vital knowledge. We urge you to visit them
all before moving on to the following more advanced resources.
For a superb, hand-holding tutorial that walks you through the entire realm
of HTML, you can’t beat W3C member Dave Raggett’s “Getting Started
With HTML” (www.w3.org/MarkUp/Guide/). It’s simple and complete; it
touches on CSS, JavaScript, and Image maps as well as HTML; and it comes
from a definitive source. (Raggett has been closely involved with the devel-
opment of HTML since the Web’s earliest days.)
The Web Design Group’s “Web Authoring FAQ” (www.htmlhelp.com/faq/

html/all.html) is yet another fine source of HTML knowledge. It even
answers questions such as “How can I get my own domain name?” “Where
can I announce my site?” and the ever-popular “How can I make a frame
with a vertical scrollbar but without a horizontal scrollbar?”
After you’ve gotten a handle on these basics, you’ll be able to learn

from the fine tutorials and articles at Webmonkey (hotwired.lycos.com/
webmonkey/), Builder.com (home.cnet.com/webbuilding/), and our own A
List Apart (www.alistapart.com). Along with technical exercises and tech-
niques, these three resources offer a bevy of useful tips, tricks, opinions,
and (best of all) insights into the changing nature of web code, design, and
content.
12 0732 CH08 4/24/01 1:22 PM Page 183
You’ll be able also to visit the W3C’s “HTML 4.01 Specification”

(www.w3.org/TR/REC-html40/) without experiencing heart palpitations.
This spec is the Mothership, though it can leave a neophyte feeling some-
what shaky. Also worth getting to know is “The Bare Bones Guide to HTML”
(www.werbach.com/barebones/) and Ron Woodall’s HTML Compendium
(www.htmlcompendium.org). These three resources provide a head-to-toe
anatomy of the HTML language. Forget an HTML tag? Not sure how one is
supposed to work? Consult these guides.
But don’t start with them. You’ll just upset yourself. HTML is simple, but
viewing dozens of pages of HTML tags and their scientific-sounding defi-
nitions can daunt the staunchest heart. Begin at the beginner’s sites, which
are written in civilian-friendly language. And learn to do one other essen-
tial thing:
VIEW SOURCE
Most of us learned HTML, not from each other’s tutorials, but by studying
the markup with which others’ web pages were built. Imitation is the sin-
cerest form of theft, and every one of us started out by copying and past-
ing other people’s markup, changing it around, and seeing what happened.
Before we discuss the ethics, here’s how to get started:
When you find a web page you like, select View, Source from your
browser’s menu bar, and save the file to your hard drive. (In Netscape,
choose View, Page Source.) Reopen it in a basic text editor (such as Sim-
ple Text or Write) or an HTML editor (such as Barebones Software’s BBEdit,
Optima Software’s PageSpinner, or Allaire’s HomeSite) and stare at the
code until it stares back. Plug your own words in between the HTML tags,
save your work, and open the file in your favorite web browser. Result: your
first (offline) web page.
Unless the layout you’ve stolen is extremely basic, you should keep it
offline. You don’t want to upload what you steal. You just want to learn
and move on. Here are the links for the editors mentioned previously:
■ Bare Bones Software BBEdit. www.bbedit.com (for Mac OS)
12 0732 CH08 4/24/01 1:22 PM Page 184
184 HOW: HTML, the Building Blocks of Life Itself: View Source
■ Optima-System Page Spinner. www.optima-system.com/

pagespinner/ (for Mac OS)
■ Allaire HomeSite. www.allaire.com/products/homesite/ (for

Windows)
A Netscape Bonus
In addition to View Source, Netscape Navigator’s browser includes a nifty
File menu option called View Page Info. Choose it, and the entire page will
be deconstructed for you in a new window, image by image. Beside each
image’s name you’ll find its complete URL (you remember…its address on
the Web), its file size, how many colors it contains, and whether or not it
uses transparency. Click the link beside each image, and the image will load
in the bottom of the window.
The way this works is not immediately intuitive; it was laid out by engi-
neers, not by designers or usability experts. But once you get the hang of
it, you’ll find View Page Info a useful tool for understanding how other peo-
ple have created web pages. IE does not offer this feature.
The Mother of All View Source Tricks

Viewing source is all well and good, but what’s even better is viewing that
source code in your text editor of choice. That way, you can continue to
work in your chosen HTML editing environment instead of dragging and
dropping (or cutting and pasting) between different software programs.
You also will have less clutter on your screen. And if your editing program
wraps text, you won’t have to cope with the endless lines of markup that
browsers often spit out by default in their built-in View Source windows.
Doin’ it in Netscape
On the Mac: Pay a visit to your Netscape Options menu. Under Options,
Applications, View Source, switch from the default program to your HTML
editing program of choice (for instance, BBEdit or PageSpinner). Now when
you view the source of any of your existing web pages, the resulting HTML
document will automatically open in your preferred HTML program, ready
for further editing. Pretty nifty. Similarly, when you View Source on some-
one else’s site, the code will open in your HTML editor of choice, tempting
you to steal other people’s work and condemn your soul to Hell.
12 0732 CH08 4/24/01 1:22 PM Page 185
In Windows: First of all, you need to install the entire Communicator pro-
gram, not just the Navigator component. From Composer (the extremely
limited semi-WYSIWYG “page creation” tool bundled with Netscape’s
browser), choose Edit then Preferences. Click Composer and register your
external editor for HTML Source. There. That really wasn’t so bad.
Doin’ it in Internet Explorer

First, open Explorer’s Preferences. Go to File Helpers and click Add.
In a new, blank dialog box, type Source Code under Description, .html
under Extension, and source/html under MIME type.
In the File Type area, click Browse. It sounds as if you’re about to browse
the Web, but you’re not. You are actually navigating your hard drive to
locate your web editor of choice. Select it, and the File type and File cre-
ator areas will be filled in automatically.
You’re not done yet. Under Handling, choose Post-Process With Applica-
tion. Hit the second Browse button, select your web editor one more time,
and then hit OK. Then stand on your head and recite the Cub Scout pledge.
Just kidding about the pledge thing.
Now when you View Source, the code will open in your favorite web edi-
tor. Not push-button easy, but it works—and you only have to do this once.
We figure these tips alone justify the cost of buying this book, and we
expect you to dog-ear this page and fondle it quietly when you think no
one is watching.
ABSOLUTELY SPEAKING, IT’S ALL RELATIVE

HTML links can work several ways. The simplest link (and often the easiest
to maintain) is the relative link.
Two files reside in the same directory:
index.html
thankyou.html
A relative link from index.html to thankyou.html looks like this:
There is a special message for you on our <a href=”thankyou.html”>Thank You</a> page.
12 0732 CH08 4/24/01 1:22 PM Page 186
186 HOW: HTML, the Building Blocks of Life Itself: Absolutely Speaking, It’s All Relative
By contrast, an absolute link might look like this:
There is a special message for you on our <a href=”http://www.ourcompany.com/

thankyou.html”>Thank You</a> page.
Or even this:
There is a special message for you on our <a href=”http://www.ourcompany.com/

customerrelations/special/thankyou.html”>Thank You</a> page.
These are called <ABSOLUTE> links because they refer to an absolute, con-
crete location in web space. (Well, as real or concrete as “web space” gets,
anyway.)
When two pages reside in the same directory, there is no need to use
absolute links. Using relative links lowers your character count (you can get
rid of http://www.ourcompany.com/customerrelations/special/), and that,
in turn, conserves bandwidth.
Relative links are easy to maintain on simple sites (though they become
fiendishly complex as a site grows and uses more and more directories). For
instance, if all images are kept in a directory called Images, the URL to an
image file might read like so:
<IMG SCR=”images/image.gif”>
We have left out the image’s height, width, and <ALT> attribute to simplify
the presentation of this idea. However, as previously mentioned, it is always
important to include an image’s height and width to help some browsers
display the layout more quickly. And, as also previously mentioned, it is
essential to include <ALT> attributes so that those with visual disabilities
or those who surf with images turned off will have some idea of the image’s
function.
The more complicated the site’s directory structure, the likelier relative
links are to require debugging. For instance, the reader is here:
somesite.com/julyissue/index.html
And you wish to direct her back to the index page at:
somesite.com/index.html
12 0732 CH08 4/24/01 1:22 PM Page 187
The URL would read as follows:
<a href=”../index.html”>Back</a> to the Index Page.
The two dots (..) preceding the slash mean “go up one directory level before
locating this file.”
With more directories, you have more and more complex links:
<a href=”../../../../../index.html”>Back</a> to the Index Page.
This can quickly lead to madness. Are you stuck writing out full, absolute
URLs? Heck, no.
Instead, you can use a shorthand form of absolute linking to retain the
advantages of relative URLs (portability, low bandwidth) while maintain-
ing the clarity of absolute URLs.
Absolute URLs also can be written like so:
/index.html
Where the slash represents “root directory.”
By using this method, if you wished to move from the July Issue index page
up one directory to the root level index page, your URL would look like this:
Return to the <a href=”/index.html”>front page</a>.
Or like this (which is even smaller and doesn’t hardcode the default
directory index filename):
Return to the <a href=”/”>front page</a>.
And reversing the direction, a link from /index.html to /julyissue/index.html

would look like this:
Read the <a href=”/julyissue/”>July issue</a>.
Unfortunately, absolute URLs of this kind cannot be tested offline. You

must load these pages to your web server to make certain the links work
correctly.
12 0732 CH08 4/24/01 1:22 PM Page 188
188 HOW: HTML, the Building Blocks of Life Itself: What Is Good Markup?
By contrast, relative links work on or offline, which enables you to keep one
or more fully functioning web sites on your hard drive.
That was relatively painless, wasn’t it? Absolutely.
WHAT IS GOOD MARKUP?

Technically, good HTML is code that validates—that is, code that fully com-
plies with current W3C standards and contains no errors. To make sure your
HTML validates, run it through the W3C validator at validator.w3.org, a free
service from those wonderful people who brought you the Web. For more
on this topic, see “HTML Standards Compliance: Why Bother?” in the Web
Developer’s Virtual Library (WDVL.com/Authoring/HTML/Standards/).
For the validator to work properly, you need to include a <DOCTYPE>. This
is a simple declaration that specifies what kind of HTML (or other markup
language) you are attempting to write. For instance:
<!DOCTYPE HTML PUBLIC “-//W3C//DTD HTML 4.01//EN”
“http://www.w3.org/TR/html4/strict.dtd”>
This declares the document to be HTML 4.01 strict. HTML 4.01 strict
emphasizes structure over presentation and balks at “deprecated elements”
such as background colors in table cells, <FONT FACE>, <FRAMES>, and
other stuff we’re supposed to do with CSS instead of in HTML.
Newer browsers such as IE5/Mac, Netscape 6, and Mozilla render HTML

4.01 strict documents according to web standards and use a “quirks” mode
for older or unspecified document types to emulate rendering bugs in older
browsers. The engineers responsible for these browsers applied these tech-
niques to offer full standards support for new sites without breaking old
sites that were written to the quirks of the companies’ older, nonstandards-
oriented browsers. Those older browsers generally ignore the <DOCTYPE>
declaration completely, but the validator requires it.
12 0732 CH08 4/24/01 1:22 PM Page 189
<!DOCTYPE HTML PUBLIC “-//W3C//DTD HTML 4.01 Transitional//EN”

“http://www.w3.org/TR/html4/loose.dtd”>
This <DOCTYPE> declares that the web page is written in HTML 4.01 tran-
sitional markup, which tolerates deprecated presentational HTML attrib-
utes (<FONT FACE>, for example) so that such documents will render
correctly in older, less standards-compliant browsers. IE5/Mac, Netscape 6,
Mozilla and IE 6 will render these documents the same way older browsers
would. This affords web designers the ability to support older and newer
browsers while making the transition from a buggy Web to one that relies
on standards. (See the section, “The 18-Month Pregnancy” in Chapter 2 to
understand why a transitional or interim period is accommodated in this
way.)
Other <DOCTYPE>s include HTML 3.2, HTML 4.01 Frameset, and XHTML
Strict and Transitional.
What Is Sensible Markup?

Conceptually, good markup is code that gets out of its own way and helps
communicate your message in the simplest, most intuitive way possible—
just like good design.
Beginning writers use too many adjectives. Beginning designers use too
many shapes, fonts, and colors. Beginning HTML authors often fall so in
love with the medium that they forget to communicate. Instead, they cram
every page with embedded MIDI (music) files, pointlessly scrolling
JavaScript messages, huge full-color photographs, animated GIFs (flames
and dripping blood are especially popular), and blinking and moving text,
often in a dozen different font faces and sizes.
That is bad design, and (we think) bad markup, even if it validates—which
is pretty unlikely because folks attracted to dripping blood animations tend
not to spend much time learning about web standards.
12 0732 CH08 4/24/01 1:22 PM Page 190
190 HOW: HTML, the Building Blocks of Life Itself: HTML as a Design Tool
Figure 8.1
This site for the Web
Standards Project contains
almost no graphics. The
shapes and colors are cre-
ated using nothing more
than HTML and CSS. It is
possible to fill the screen
with color and content
without wasting band-
width on images. As a
bonus, the code validates
(www.webstandards.org).
HTML AS A DESIGN TOOL

Though this won’t always be the case, one of the beautiful things about
HTML (and eventually, CSS instead) is that it can be used as a powerful
design tool—a design tool that loads instantly. No images are required;
there are no fancy plug-ins and no worries about every user having the lat-
est browser.
Consider the front page of The Web Standards Project (www.webstandards.

org). Aside from one large Seymour-Chwast-like illustration, the rest of the
front page is designed entirely with HTML and CSS. Now view the source.
An HTML color in the <BODY> tag defines the entire background. The con-
tent grid is made up of a table, and the grid areas and background colors
are defined with table cell colors.
12 0732 CH08 4/24/01 1:22 PM Page 191
The content area is enclosed within a black outline created with one line
of CSS. Originally, the same effect was created by wrapping one HTML
table inside another.
CSS is used to create the typography and leading.
Creating a layout like this in Photoshop, cutting its elements into pieces,
and assembling those pieces via HTML, would have resulted in a large web
page composed of many small files that would take a long time to down-
load. (And if the visitor had images turned off, the visual effect would be
lost.)
Creating the layout in HTML and CSS means that the page loads almost
instantly, no matter how slow the visitor’s connection may be. And the lay-
out is backward compatible with browsers dating back to Netscape 3 (a
1997 browser), although the CSS formatting will be lost in that browser.
Actually, the site is viewable in any browser. Older browsers will lose the
design branding effects, but the content will still be readily accessible.
Note that this is a transitional web design strategy. It respects bandwidth

by using web technology (instead of image files) to create visual and
branding effects—but at the cost of relying on deprecated presentational
HTML attributes. Most of these effects can be generated in CSS alone, with
HTML serving simply as the structural container of content. This is what
the W3C recommends, and this is the way we will build all sites in the near
future and forever after. (We’re doing it at alistapart.com.)
However, as we mentioned in Chapter 2, old browsers that do a poor job at

understanding CSS are still widely used in the market we serve. And as
you’ll see in Chapter 10, browsers that stumble over CSS don’t simply ren-
der it incorrectly. They can actually crash and burn. For this reason, as you
begin your career in web design, you will undoubtedly be using HTML tables
and other deprecated presentational HTML attributes to control your web
layouts and visual effects. Thus there is value in learning how to do this in
ways that minimize wasted bandwidth and comply with the letter (though
not the spirit) of W3C standards. The lessons you learn in building sites this
way will apply equally well when you are free to control your site designs
exclusively with CSS.
12 0732 CH08 4/24/01 1:22 PM Page 192
192 HOW: HTML, the Building Blocks of Life Itself: HTML as a Design Tool
Other sites that use HTML or XHTML as a creative design tool (abetted
by CSS):
■ www.glish.com (designed by Eric Costello)
■ www.harrumph.com (designed by Heather Champ)
■ www.assembler.org (designed by Brent Gustaffson)
■ www.kottke.org (designed by Jason Kottke)
■ www.blogger.com (designed by Derek Powazek)
■ a.jaundicedeye.com/weblog/ (designed by Steven Champeon)
■ www.zeldman.com (designed by Zeldman)
■ www.alistapart.com (designed by Zeldman)
Use View Source to see how these sites use HTML table cells and table cell
colors, CSS fonts, leading, margins, and background images to create full-
Figure 8.2
This site, though colorful,
contains few graphics.
Big color sections are
created with CSS and
HTML <DIV>s. The tech-
nique facilitates Liquid
Design, reduces band-
width, and makes the
site more accessible
(www.alistapart.com).
12 0732 CH08 4/24/01 1:22 PM Page 193
fledged visual experiences using nothing more than code (and a few low-
bandwidth images).
PLUG-INS AND TABLES AND FRAMES,

OH MY!
In the transitional Web, designers use HTML tables to lay out pages, as just
described (with additional commentary and how-to-do-it type verbiage to
come in the next chapter). We also use <FRAMES>, a Netscape “extension”
to HTML which has temporarily made it into the HTML 4 Transitional stan-
dard but which will eventually go the way of the Dodo bird.
Frames are nothing more than pages within pages, for example:
<FRAMESET COLS=”80,2,*” frameborder=”no” border=”0” framespacing=”0”>
<FRAME SRC=”nav.html” NAME=”nav” marginwidth=”0” marginheight=”0”
noresize scrolling=”auto”>
<FRAME SRC=”black.html” NAME=”black” marginwidth=”0” marginheight=”0” noresize
scrolling=”no”>
<FRAME SCROLLING=auto SRC=”content.html” NAME=”content” marginwidth=”0”
marginheight=”0”>
</FRAMESET>
In this markup, <FRAMESET> tells the browser that the page contains
frames. <COLS> (short for columns) specifies that the frameset contains
three columns. The first is 80 pixels wide; the second is 2 pixels wide; and
the last fills the remaining width of the browser window.
We also can tell the browser whether or not we want borders on our
frames; whether or not each frame should permit the viewer to scroll con-
tent; whether or not each frame is user-resizable; and what size margin
we’d like on each frame. Because we’re designers, we turn margins off
entirely (marginwidth=”0” marginheight=”0”) and use CSS or tables to
control the margins on each individual frame—each frame, of course, being
nothing more than an HTML document (nav.html, black.html, content.
html).
We also name each frame for targeting purposes. After all, when visitors
click in a menu area, we want the content they’ve chosen to show up in
the content frame—not in the menu frame. Assigning a target name to each
12 0732 CH08 4/24/01 1:22 PM Page 194
194 HOW: HTML, the Building Blocks of Life Itself: Plug-ins and Tables and Frames, Oh My!
frame enables us to write links like <a href=”companypolicy.html” target

=”content”>. Naming also enables us to perform JavaScript stunts, which
are mainly useful for avoiding the maddening usability hazards associated
with frames, such as the following.
The Frames of Hazard

Through a search engine such as Google.com, Aunt Moira finds one of our
client’s content pages. Unfortunately, that’s all she finds—the naked con-
tent page, immodestly lacking its associated navigational menu frame.
Aunt Moira has no idea where she is, where she can go next, who created
the site, or even how to find its home page.
We always had it in for Aunt Moira, who never failed to point out when we
had gained weight or burst out in pimples, but our client would like her to
be able to use the site. Because we have named our frames and because
the good Lord (well, actually, Netscape) gave us JavaScript, we can instruct
the browser to load named frames if they are not already visible on the
page.
Thus when Aunt Moira (the old biddy) blunders her way into
companypolicy.html, the browser is instructed to load the missing frames.
Code like this would appear on each HTML document that made up the
frame, though not on the frameset document itself:

<script LANGUAGE=”JavaScript”>

</script>
Such a script tells the browser to make sure that frameset.html has loaded.
If it has not loaded—if Aunt Moira’s browser is about to show a confusing
single frame no more illuminating than a single puzzle piece—the browser
gathers and assembles the missing pieces before the dear old thing has a
chance to notice that anything is amiss. We don’t want to get ahead of
ourselves by discussing JavaScript in the HTML chapter. Suffice to say, the
need to rely on such scripts points out some of the hazards of HTML
<FRAMES>, and helps explain why they are on their way out.
12 0732 CH08 4/24/01 1:22 PM Page 195
Everything we now do with frames (and more) we can do with CSS, which
is standards-compliant and avoids the usability and accessibility hazards
<FRAMES> engender. But to switch from <FRAMES> to CSS, we must wait
for some browsers to improve their CSS support and all users to upgrade
to these better browsers.
Please Frame Safely

Some old browsers do not understand frames. Neither do text and audio
browsers. Nor do Palm Pilots and web-enabled telephones. To accommo-
date these devices and browsers, your frameset should include a
<NOFRAMES> tag. Browsers that can’t read frames can read the plain
HTML that you insert between opening <NOFRAMES> and closing
</NOFRAMES> tags. Copy the content that appears in your frames, paste
it between the <NOFRAMES> tags, and you are on your way to creating a
site everyone can access, regardless of their browser’s capabilities.
Framing Your Art

Despite these hazards and hassles, frames can be quite useful to web
designers. Frames allow you to present a menu bar that stays in place while
content frames change. They also enable you to create layouts where, for
instance, your content will always appear in the center of the screen,
regardless of the visitor’s monitor size. View Marc Klein’s Creative Repub-
lic (www.creative-republic.com) to see this in action, and then view the
source to see how Marc crafted his framesets.
One other maddening thing about frames is that though Netscape invented
them, it never got them exactly right. When you tell the browser to make
your menu bar 25 pixels tall, you may get 25 pixels in Netscape 4, but
you’re just as likely to get 32 or 16. If this suggests that you’re better off
avoiding frames whenever possible, we won’t argue.
In addition to tables and frames, web designers use applets and multime-
dia files to create designs that are frankly unimaginable in print. We will
discuss those in Chapter 12, “Beyond Text/Pictures.” Don’t skip ahead, we’re
watching.
12 0732 CH08 4/24/01 1:22 PM Page 196
196 HOW: HTML, the Building Blocks of Life Itself: Plug-ins and Tables and Frames, Oh My!
Figure 8.3
Creatively used frames
keep design elements
fixed in the center of the
screen, whether the visi-
tor’s monitor is large…
Figure 8.4
…or small. Designer:
Marc Klein
(www.creativerepublic.com).
12 0732 CH08 4/24/01 1:22 PM Page 197
<META> <META> HINEY HO!

Though <META> tags have many purposes, web designers and developers
most often use them for one of two reasons:
■ Accommodating search engines
■ Reloading pages or forwarding visitors to an updated page
Regardless of the application, <META> tags are placed in the <HEAD> sec-
tion of HTML markup. That is, all <META> elements show up between the
<HEAD> and the </HEAD> tags. Now let’s wrap our own <HEADS> around
them to see how this all works:
Search Me
When Aunt Moira (the old battleaxe) enlists the help of a search engine to
find a topic or subject, one way in which the search engine might sort data
is through <META> tags. Some search engines compare search words with
<META> descriptions, and they return the web pages that provide the best
matches, as in the following:
<META NAME= “author” CONTENT= “your name”>

<META NAME= “description” CONTENT= “page description”>
<META NAME= “keywords” CONTENT= “keywords that apply to your page”>
<META NAME= “generator” CONTENT= “the editor you used to create your page”>
<META NAME= “copyright” CONTENT= “date of copyright”>
<META NAME= “expires” CONTENT= “expiration date”>
Most corporate and business-to-business sites will include only the

<DESCRIPTION>, <KEYWORD>, and <COPYRIGHT> tags. After all, AT&T
does not need its customers to know who designed the site, what tool they
used to edit the HTML, or how old (and outdated) the page may be.
Aside from <HTTP-EQUIV> (the widely accepted predecessor to <DOC-

TYPE>), there is no reliable standard for <META>. Most search engines
rarely use them (Google, for instance, ignores them). Those such as
Altavista and Hotbot, which once relied on them extensively, pay them less
and less heed as time goes by. Good <TITLE> tags and good, descriptive
page copy are more effective at scoring with search engines and
directories.
12 0732 CH08 4/24/01 1:22 PM Page 198
198 HOW: HTML, the Building Blocks of Life Itself: <META> <META> Hiney Ho!
In spite of everything we’ve said, some search engines and directories do

pay attention to these tags, and it sometimes falls to the designer to write
them. So let’s look at some good and bad ones. Here is a good one:
<META NAME= “description” CONTENT= “Widgets.com builds reliable widgets for the
lubrication industry. As the American Midwest’s largest developer and supplier of indus-
trial-strength widgetry, we offer a product line of 2,000 parts as well as custom products
built to your specifications. Standard products ship in 48 hours in the Continental U.S., and
within three business days to lubricant concerns in Europe, Asia, and Africa.”>
<META NAME= “keywords” CONTENT= “widgets, lubricants, lubrication, industry, U.S., mid-
west, developer, supplier, industrial strength, widgetry, 2,000, standard, parts, custom,
product, development, shipping, 48 hours, Canda, Europe, Asia, Africa”>
<META NAME= “copyright” CONTENT= “12 January 2001”>
And here is a bad use of <META> tags:
<META NAME= “description” CONTENT= “Welcome to our home page on the World Wide
Web! We are happy to serve you. Please do not hesitate to call on our reliable staff if we
may serve you better in any way, shape, or form. This site is under construction. Some links
may not work and some pages that we are going to make later have not shown up yet
because we are still arguing about them in the boardroom. All our products are proudly
made in the good old U.S. of A. We are a good company that has serious social concerns.
Kids, stay in school. Hugs, not drugs. Have a nice day.”>
<META NAME= “keywords” CONTENT= “welcome, to, our, home, page, which, is, under,
construction, serving, you, proudly, since, 1955, but, not, the, website, which, as, we, men-
tioned, is, under, construction”>
<META NAME= “author” CONTENT= “your name here”>
<META NAME= “generator” CONTENT= “Hot Dog Pro”>
The good <META> tags help search engines hone in on what the site actu-
ally has to offer. The bad <META> tags consign the site to the dung heap,
where it will never be found by any living soul—unless they are searching
for serve + kids + drugs.
Raw-elbowed marketing idiots, who are legion on the Internet, used to try
to “upgrade” their search engine rankings by repeating certain keywords—
a practice referred to as keyword spamming.
<META NAME= “keywords” CONTENT= “widgets, widgets, widgets, widgets, widgets,

widgets, widgets, widgets, widgets, widgets, widgets, widgets, widgets, widgets, widgets,
12 0732 CH08 4/24/01 1:22 PM Page 199
widgets, widgets, widgets, widgets, widgets, widgets, widgets, widgets, widgets”>
Needless to say, this no longer works, and if anything, you and the widg-
ets you rode in on will be dropped to the very bottom of any halfway rel-
evant search—or kicked out of the database altogether. Kids, don’t try this
at home (page).
As we say, most search engines ignore <META> tags, so if you want your
site to be found, focus on developing relevant body text and <TITLE> tags.
“Welcome to our home page on the World Wide Web” is not relevant text.
“Widgets.com builds reliable widgets for the lubrication industry” is rele-
vant, if unsavory, text. <TITLE> tags and body text are weighted more heav-
ily than <META> tags, even by search engines that consider all three
(<META> tags, <TITLE> tags, and body text). This is because it is easy for
liars to lard their <META> tags with exciting buzzwords that have little to
do with what the site actually offers. Body text—text seen by visitors—is
therefore given precedence over the wishful thinking that goes on inside
the <META> tag.
Everything we’ve just told you is probably outdated and irrelevant by now.
Visit www.searchenginewatch.com to get the latest specifics on search
engine ranking.
At a bad shop, <META> tags (and indeed, sometimes, body text) will be
written at the last minute by a recent college graduate with no experience
in marketing, communications, or the Web. When shopping for a job, don’t
simply judge the company by its graphic design. Peek under the hood for
evidence of a caring, intelligent environment—or a sweatshop that bangs
work out with little regard for its success or failure in the marketplace.
Wow, we’ve just saved you from taking a really bad job. This book is turn-
ing out to be worth every penny you paid for it, isn’t it? You ought to buy
copies for all your friends, and save them from taking bad jobs, too.
Another type of <META> tag (the <META HTTP-EQUIV>) does another type
of job and is worth mentioning.
12 0732 CH08 4/24/01 1:22 PM Page 200
200 HOW: HTML, the Building Blocks of Life Itself: A Comment About <COMMENTS>
Take a (Re)Load Off

There are times where you want a page to hesitate and then reload. Though
this may sound like a tricky process, <META HTTP-EQUIV> tags make it
barnyard-simple:
<META HTTP-EQUIV= “REFRESH” CONTENT= “x; URL=http://www.widgets.com/”>
In this code example, <x> represents the number of seconds before the
refresh or reload occurs, and the URL refers to the page currently being
viewed. (Obviously, you would replace <x> with <10>, <6>, or however
many seconds you wish to have elapse before the page reloads itself. There
is no limit, to our knowledge, on how many seconds that may be. The
browser tells time via the operating system. Uncanny, is it not?) Given that
the visitor is already at www.widgets.com, why spell out the full URL
instead of a relative URL (such as index.html)? Trust us on this one. (If you
don’t trust us, using a relative URL will usually work, but can be problem-
atic if the page you’re refreshing gets moved or renamed, which web pages
often do. Full URLs make for better, safer maintenance in this instance.)
You also can use this technique to forward the visitor from an old, outdated
page to a shiny new one:
<META HTTP-EQUIV= “REFRESH” CONTENT= “x;

URL=http://www.widgets.com/newindex.html”>
Many HTML experts, being spoilsports who live in Ivory Towers and proba-
bly never laugh even at really funny jokes like the one about the traveling
salesman, the farmer, the hippie, and the bus driver, disapprove of this
entire procedure. They recommend that you forward web users to new
pages (if need be) by using JavaScript. However, this <META> tag tech-
nique does work, even with old, non-JavaScript-capable browsers.
A COMMENT ABOUT <COMMENTS>

In your career as a web designer, you will sometimes create entire sites by
yourself from scratch. Most of the time, though, you will be working with
a team. Occasionally, you will inherit an existing site that needs to be
12 0732 CH08 4/24/01 1:22 PM Page 201
redesigned or updated. At other times, you will be creating a site for some-
one else to update. All these situations are best served if you comment the
code as you write it. Referring once again to the code used earlier in this
chapter:


<table border=”0” cellpadding=”0” cellspacing=”0” align=”center”>
<tr>
<td>
<a href=”reading.html”><img src=”reading.gif” width=”20” height=”20” border=”0”
alt=”Reading”></a>
</td>
<td>
<a href=”writing.html”><img src=”writing.gif” width=”20” height=”20” border=”0”
alt=”Writing”></a>
</td>
<td>
<a href=”arithmetic.html”><img src=”arithmetic.gif” width=”20” height=”20” border=”0”
alt=”Arithmetic”></a>
</td>
</tr>
</table>

<Begin menu bar> and <End menu bar> are the comments that help you
(or a teammate or successor) figure out what was intended by all that
wacky HTML. They are always enclosed within  so
that they will not be displayed on the web page. Even if you routinely work
alone (say, as a freelancer), comments will help you find your way when
you return to an HTML document you haven’t looked at for six months. Pro-
fessional web designers always comment their markup.
In Chapter 2 we mentioned that designers could save bandwidth by remov-

ing white space from their HTML documents. We also mentioned that most
of us refrain from this practice because it interferes with the need to con-
tinually update existing web documents. Comments exist to facilitate that
need. No further comment.
12 0732 CH08 4/24/01 1:22 PM Page 202
202 HOW: HTML, the Building Blocks of Life Itself: WYSIWYG, My Aunt Moira’s Left Foot
WYSIWYG, MY AUNT MOIRA’S LEFT FOOT

We’ve all seen the ads: “Create web pages without learning a single HTML
tag!” We’ve also seen ads that tell us how to lose weight while eating candy
bars all day long. Strangely enough, we know no one who’s lost weight that
way.
Today’s “What You See Is What You Get”( WYSIWYG) programs are far more
powerful than the early, lame-o programs that gave WYSIWYG a bad name.
But most professional web designers continue to use text-based web edi-
tors. Why? In a word, control. In four words, to avoid bad markup.
Code of Dishonor
Though we hope to see this change soon, nearly all WYSIWYG editors tend
to write bloated (and often invalid) HTML markup. To make sure that every
browser—even one that’s five years old—will be able to display your page
as the program thinks you want it to be seen, these programs will grind out
all kinds of unnecessary workaround markup, adding unsightly flab to every
web page.
Other programs, notably one famous one we won’t mention for fear of law-
suits, tend to generate markup that works only in one browser. Coinciden-
tally, this browser is made by the company that also makes the WYSIWYG
program. Is this just bad design or an insidious marketing ploy? Ask their
attorneys.
Beyond the twin plagues of page-swelling bloat and browser-specific

“HTML,” there is the problem of artificial limitations imposed upon you by
the designers of any WYSIWYG program you may use. Unless you work the
code yourself, you cannot expand its capabilities or explore new creative
terrain.
Citizen Kane was not shot with an autofocus lens. Great web pages are not
built by using defaults. Use the markup, or you’ll be forced to depend on
the kindness of strangers (otherwise known as software companies), to
determine what you can and cannot do with your site.
12 0732 CH08 4/24/01 1:22 PM Page 203
With an autofocus camera, the man in the striped hat will be in perfect
focus; too bad if you wanted to focus on the bird in the bush. Likewise, even
with an advanced WYSIWYG editor, your options as a designer will always
be limited. Comparing WYSIWYG editors to autofocus cameras is probably
unfair—to the cameras.
Yes, these WYSIWYG programs are getting much better. Yes, a substantial
number of pros do use them, particularly to rough out web pages quickly.
But these pros always end up revising the end product by hand.
WYS Is Not Necessarily WYG

With a WYSIWYG tool, if you slap an image down 30 pixels to the right of
another image, it stays 30 pixels away, even if you want it to move as the
user’s window widens. If you drop an image onto the exact center of the
WYSIWYG editor page, you might think the image is “centered,” but it’s
not—it is stuck in an exact location, which may bear no relation whatso-
ever to the relative center of your users’ respective browser windows. (This
is also the problem with using more advanced WYSIWYG editors to gener-
ate DHTML pages or CSS-based layouts. But we’ll get to those issues in
time.)
WYSIWYG editors give you a false sense of control and a false sense of the
Web. As explained in Chapter 2, the Web is not fixed like a printed page. It
is fluid and variable and should be designed for accordingly. The tightly-
rendered page that looks great in your WYSIWYG editor may look terrible
on Aunt Moira’s monitor because your default fonts are larger than hers,
or she doesn’t have the same fonts installed that you do, or just because
she’s a silly thing who is going to leave her money to her cats, not you.
Suppose we intend to create a three-column layout with an image in the

center column. Using HTML, this is no problem—we write a three-column
table, set its borders to 0, and in a few moments, we are done. If we’ve used
relative widths when constructing our table (<width=”33%”> for example,
instead of <width=”200”>) the design will reflow to accommodate any
user’s monitor, as discussed back in Chapter 2.
We can do the same thing with CSS, and before this book reaches its sec-
ond edition, that’s what we’ll all be doing. With CSS such layouts are faster
and easier to achieve, and the resulting web pages render more quickly.
12 0732 CH08 4/24/01 1:22 PM Page 204
204 HOW: HTML, the Building Blocks of Life Itself: Browser Incompatibilities
Now let’s build the same layout in a WYSIWYG editor. We drag three
columns over a grid and place our image in the middle column. Unfortu-
nately, we were two pixels off when we dropped our image, because the
program lacks a “snap-to-grid” feature (or we forgot to turn the feature
on). What does the program do? It calculates an 18-column cubist mess of
code, using <ROWSPANS> and <COLSPANS> to make sure that our mis-
take gets perfectly rendered.
The program doesn’t know that our inexact placement of the image was an
accident. The program cannot think; it can only execute, using tortured
workarounds to honor our errors as hidden intentions. The result is a slow-
to-download, tortuously coded fiasco—one which, after all that absurd
markup and lengthy downloading, looks like garbage because the layout is
subtly “off.”
And of course, it will never reflow to fit each user’s monitor just so.
Knowing HTML doesn’t make you a web designer any more than knowing
your native language makes you a writer. But choosing not to know is
senseless. Don’t trust the ads. Learn the markup. If you wish to use the bet-
ter WYSIWYG programs to rough out your layouts, go ahead, but be ready
to get in there later and refine your code.
BROWSER INCOMPATIBILITIES: CAN’T WE

ALL JUST GET ALONG?
Not only is there no WSY in WYSIWYG web editors, there’s no guarantee
that any two browsers will display your page the same way or even that
your page will work in every browser. Even if you write perfectly valid and
standards-compliant code, old browsers are not standards-compliant, and
the dream of “write once, publish everywhere” has not yet been attained.
Moreover, even on that great day when all browsers fully support W3C
standards, extensive platform and hardware differences (as described
extensively in Chapter 2) mean that the Web will remain evanescent and
unfixed: a little different with each browser, in each monitor, and on each
operating system. That kind of incompatibility is perfectly okay—there’s
nothing we can do about it anyway. Incompatibilities that result in page
failures are not okay.
12 0732 CH08 4/24/01 1:22 PM Page 205
One thing you can do is author in accordance with commonly supported

web standards instead of to “nifty new features” that work only in one
browser. By definition, you will be including more people if you avoid pro-
prietary, browser-specific markup. Given that support for these standards
varies widely and browsers may legitimately differ in the way they inter-
pret some standards, you and your company’s Quality Assurance (QA) team
will spend much time testing designs on a variety of browsers and
platforms. (See Chapter 7, “Riding the Project Life Cycle,” if you skipped it
earlier.)
Another thing you can do is visit The Web Standards Project (www.
webstandards.org), read our Mission Statement (www.webstandards.org/
mission.html), and use the Project’s Resources section to learn more about
standards (as well as incompatibilities). (In Chapter 10 we’ll talk about CSS
incompatibilities and how to work around them.)
PUBLISH THAT SUCKER!

After you have created a website, how do you publish it? You publish it by
sending your files and directories to the web server. This is done by means
of an FTP program, so called because it uses the File Transfer Protocol (FTP)
to do its work. Fetch is one common FTP program for the Mac; Interarchy
(the FTP program formerly known as Anarchie) is another; and Panic Soft-
ware’s Transmit (www.panic.com/transmit/) is a third—and the most Mac-
like. We still use Fetch, which has not been updated since the Pleistocene
era, because the crusty old tool makes us feel that we are in UNIX, and that
makes us feel all hardcore and stuff. WinFTP and CuteFTP are common
Windows FTP programs.
To use an FTP program, you open it, type in the FTP address, user name, and
password, and upload your files by dragging them from the open window
on your desktop to the open FTP window. You can drag and drop hundreds
or even thousands of files at once.
Note that unlike the Mac OS, an FTP server will not warn you if you are
about to overwrite your files. Nor is there a comforting “Are you sure?” dia-
log box, such as in Windows. (Well, maybe the “Are you sure?” box is not
12 0732 CH08 4/24/01 1:22 PM Page 206
206 HOW: HTML, the Building Blocks of Life Itself: Publish That Sucker!
comforting, exactly, but it does help prevent mistakes. FTP does not.) Exist-
ing files, if present, will simply be deleted and replaced by the new file.
Many a life, or at least, a weekend, has been ruined when a web designer
dragged one file on top of another. So use care when naming
your files. Many web designers rename old files before they update
them (personnel.html becomes, for instance, personnelbak.html, or
~personnel.html).
Equally important is that depending on the rules of the FTP server, text files
might have to be uploaded as text, or they will not work. Image files, along
with Flash movies, sound files, and so on, might have to be uploaded as
binaries, or they will not work. Doddering old Fetch has a checkbox for
“automatic” detection of text or binary. That checkbox is your friend. Check
it and you will not be faced with the mysteries of the nonworking site.
Finally, as we’ve emphasized all along, it’s important to make sure that your
files end in appropriate extensions (.jpg for JPEG images, .html for HTML
documents, and so on) and that you have paid attention to their capital-
ization—or lack thereof.
Offline, you can get away with mismatched cases. For example, <IMG
SRC=”mydog.gif”> might work just as well as <IMG SRC=”MYDOG.GIF”>
or <IMG SRC=”mYdOg.gIf”> when you’re testing the web page offline on
your hard drive. But almost all web servers are case-sensitive. (Windows IIS
does not seem to care one way or the other.) On most servers, if the file is
named mydog.gif and your HTML refers to <MyDog.gif>, the image will not
show up on the Web.
Many web designers avoid this problem by using only lowercase for their
filenames: mydog.gif—never MyDog.gif or MYDOG.GIF.
Sticking to lowercase and coding all references in lowercase may save

hours of tedious labor. You’ll also protect your clients and your site’s poten-
tial visitors. Because most folks who’ve spent time on the Web have noticed
(consciously or unconsciously) that nearly all URLs are lowercase, when
they hear your client’s ad they’ll type http://www.widgets.com. They will
not type HTTP://WWW.WIDGETS.COM. Stick to lowercase so your client’s
visitors can actually view the site.
12 0732 CH08 4/24/01 1:22 PM Page 207
Besides, all-caps filenames are annoying. Who wants to view MYDOG.GIF

on MYHOMEPAGE.HTML? Come to think of it, who wants to view mydog.gif
on myhomepage.html? Never mind.
One of our clients performs his own site maintenance and updating. Well,
actually, many of our clients do this, but we’re not talking about those
clients. We’re talking about a particular client who wreaked havoc by
renaming a certain directory <PRODUCTS> after linking to it throughout
the site from its original name, <products>. One little word, eight little let-
ters that simply meant he got fired.
HTMHELL
This chapter and the resources to which it points are not sexy because
HTML is not sexy. It is a dull, baseline standard that behaves in predictable
ways (give or take a few browser compatibility problems). As a web
designer, you’ll be hired because of your visual skills and your thinking, not
because you can upload files correctly, write good <META> tags, or have
committed the various <DOCTYPES> to memory. Nevertheless, without a
thorough understanding of HTML and the ability to write it, detect and fix
errors in it, and use it creatively as a design tool, you cannot be an effec-
tive web designer. So take the time to learn this simple, logical markup lan-
guage before moving on to the more exciting stuff. (The exciting stuff
begins in the very next chapter.)
12 0732 CH08 4/24/01 1:22 PM Page 208
13 0732 CH09 4/24/01 11:21 AM Page 209
chapter 9
Visual Tools
IN THIS CHAPTER, you’ll learn how web designers use Adobe Photoshop and
related software to design comps, prepare typography and images, and
convert the whole shebang into working web pages. Along the way, you’ll
get the lowdown on image file types, learn design techniques that make a
virtue of web images’ limitations, and see how the issues of color, band-
width, and navigation discussed earlier in this book apply to the creation
of web layouts in image editors. We’ll also chat about alternative web
design methods that produce lighter, more accessible sites.
If you’ve read other web design books, some of the initial material in this
chapter will be familiar to you, though we might take it places other books
haven’t.
In short—pour yourself a tall one, fluff up your seat cushions, and get ready
to burrow in.
PHOTOSHOP BASICS: AN OVERVIEW

Coming from the world of print, most art directors and designers are famil-
iar with Adobe Photoshop as an image editing tool. In web design, Photo-
shop is that and more. In fact, Photoshop, along with its included
ImageReady module, is most web designers’ primary imaging, layout, and
production tool.
13 0732 CH09 4/24/01 11:21 AM Page 210
210 HOW: Visual Tools: Photoshop Basics
Some web designers use Macromedia Fireworks (www.macromedia.com/

software/fireworks/) to supplement or even replace Photoshop. Fireworks
is a fine tool created specifically to serve the needs of web design. But as
a transitioning designer or as one adding web work to an existing reper-
toire of design services, you will want to use the tools you know. And that
means Photoshop/ImageReady and Illustrator. You will encounter Fire-
works in some web agencies—Photoshop and Illustrator in all of them.
We’re assuming that you already know how to open an image in Photo-
shop, resize it as necessary, apply color correction, make selections, run fil-
ters, save the image in a particular format, and scream when the client tells
you your multilayered masterpiece is “too busy.” If not, now might be a
good time to brush up on your basic Photoshop skills (www.adobe.com/
products/tips/photoshop.html).
Following is an overview of key Photoshop functions in addition to the

familiar tasks of resizing, color correction, blurring, and sharpening. Mate-
rial that might be new to you will be covered in detail following the
overview.
Comp Preparation
Unlike in the print world, where Quark XPress, Illustrator, and InDesign hold
sway, most web designers create their page layouts entirely in Photoshop.
You’ll use it to conceive designs and show them to clients.
Dealing with Color Palettes

In print, color is practically unlimited. Not so on the Web. Photoshop 5.5
(or higher) and its bundled sister product, ImageReady, handle this issue
with ease and grace.
Exporting to Web-Friendly Formats

Each computing platform sports a native, bitmapped image format—PICT
for Mac users and BMP for Windows. But web browsers are configured to
display special, cross-platform image formats that trade quality for band-
width. In designing web pages, you’ll use the compressed GIF and JPEG for-
mats almost exclusively. The PNG format, an open standard with
13 0732 CH09 4/24/01 11:21 AM Page 211
advantages including alpha channel transparency, is also beginning to

enjoy support in newer browsers. Photoshop exports to all these formats,
with advanced functions that make your job easier. It is also a fine tool for
applying image compression during the exporting process.
Gamma Compensation
Photoshop easily handles the cross-platform gamma dilemma we dis-
cussed earlier in this book. (See “Gamma, Gamma, Hey!” in Chapter 2,
“Designing for the Medium.”)
Preparing Typography
Photoshop, together with Illustrator, enables you to prepare typographic
images for the Web. Photoshop has become so adept at this task that many
web designers now use it exclusively.
Slicing and Dicing

To turn a comp into a web page, most professionals find themselves slic-
ing the comp into smaller component images and using HTML markup to
put the pieces back together. Photoshop and ImageReady make this easy
and painless, relieving you of the burden of hand-coding complexly nested
HTML table cells and their associated image files.
Rollovers (Image Swapping)

The ever-popular rollover effect, in which one image is replaced by another
when the visitor’s cursor “rolls” over it, is not just a meaningless gimmick.
By emulating familiar Graphical User Interface (GUI) behavior, in which
user actions trigger software reactions, rollovers can provide important
cues to the way the site functions. Or they can just be meaningless gim-
micks. Rollover effects are powered by JavaScript (or ECMAScript, as it now
prefers to be called).
We’ll explore JavaScript in Chapter 11, “The Joy of JavaScript.” While there
is no substitute for learning JavaScript and employing it creatively, in this
chapter you’ll learn how ImageReady can automatically generate appro-
priate rollover scripts for you. These rollovers can be extremely sophisti-
cated and might exceed many web designers’ hand-programming abilities.
13 0732 CH09 4/24/01 11:21 AM Page 212
212 HOW: Visual Tools: Color My Web
GIF Animation
On the Web, images need not be static. Animated GIFs create the illusion
of motion without requiring visitors to download and install third-party
add-ons such as Flash, Shockwave, or the Adobe SVG plug-in (not that
there’s anything wrong with Flash, Shockwave, or SVG, all of which are dis-
cussed in Chapter 12, “Beyond Text/Pictures”).
GIFs can contain more than one image, and the format was originally
prized for its utility as a kind of multiple image storehouse. In the mid-
1990s, some smart soul figured out that these multiple images could be
“played” in sequence, creating the illusion of motion. The animated GIF was
born, and the Web has never fully recovered. Photoshop’s ImageReady
module enables you to easily create GIF animations. These can be free-
standing, but might just as easily be incorporated into rollovers.
Create Seamless Background Patterns (Tiles)

These patterns or tiles formed a staple of web design in its early years.
Many were downright ugly, and few appear in today’s sophisticated sites,
but the technique can still prove useful when creatively reimagined by web
designers with taste.
From this brief overview, it should be clear that the Photoshop/ImageReady

combo is a powerful tool for web designers. Basically, with Photoshop and
your HTML editor of choice, you can perform almost any web task.
Now let’s look at some problems peculiar to web design and see how you
can solve them with Photoshop and ImageReady.
COLOR MY WEB: ROMANCING THE CUBE

Glance back at Chapter 2 for a refresher on the 216 color palette—or the
Netscape Color Cube.
Designers work with computers that support millions of colors. But most
web users are limited to thousands (or hundreds) of colors, and your design
must work well in these environments.
13 0732 CH09 4/24/01 11:21 AM Page 213
Monitors limited to thousands of colors (16 bits) might seem to display

realistic color, but it is never the actual color specified by the web designer.
For mathematical reasons, colors shift slightly “off” in the 16-bit color
space. This problem is insoluble and will haunt you like Jacob Marley’s
ghost until cheap 24-bit graphics cards find their way into most PCs and
vendors ship them configured to use the higher resolution and bit depth.
(One of the tragic stupidities of the computer industry is that computers
that can display millions of colors come configured to show thousands;
those that can show thousands come configured to show hundreds, and so
on. It’s tragic because ordinary citizens rarely realize that they can increase
their PC’s graphic power with a quick trip to the appropriate control panel.)
Eight-bit (256 color) systems face an additional problem in that up to 40

of these 256 colors are “used up” in advance by the operating system itself.
For instance, Windows reserves 40 (count ‘em) Windows system colors for
its own display purposes. Knowing Windows, we should be glad it’s only 40.
Nevertheless, that leaves exactly 216 colors at your disposal. (And GIF, as
an “indexed” file format, only supports 255 colors anyway, two of which—
black and white—are always present.)
What happens to viewers with lower-end graphics capabilities when you

design with millions of colors they can’t see? The browser tries to simulate
your color choices by combining adjacent pixels of color the visitor can see.
This visual side effect is known as dithering, a verb that also means “bab-
bling inconsequentially,” which is kind of what we’re doing here.
Dither Me This
You’ve chosen a subtle shade of off-white for your typography. The viewer’s
graphics processor cannot reproduce that exact color, so the web browser
breaks up your type into a series of adjoining pink and white pixels. If the
viewer squints, she will get an approximation of the color you intended to
use (see Figure 9.1).
13 0732 CH09 4/24/01 11:22 AM Page 214
Figure 9.1
The toothpaste may get
teeth their whitest, but
it doesn’t do much for
this off-white typographic
headline. On 8-bit
systems, the type gets
pixellated, and we suspect
the web designer will, too.
(Image enlarged 200%.) In small, transitional areas, dithering is okay. But when it occurs across
large areas of solid color—or when it is visible in the primary letterforms of
typography—the result will be visually hideous, and legibility can be seri-
ously impaired. (Usability experts and web artists alike can agree that
hideous, illegible type is not a good thing.)
Because the discrepancy between computers’ graphic capabilities is so

enormous, it initially seems as though it would be impossible for a designer
to create web pages that do not dither and degrade on most viewers’ mon-
itors. The Color Cube saves the day (see Figure 9.2).
Figure 9.2
With the typography
recast in web-safe white
(#ffffff), the headline is
no longer pixellated,
increasing the chances
that it will actually be
read. The background
image is still dithered, but
users of 8-bit systems will For typography, CSS or HTML background colors, or any other area of large,
accept that. (Image flat color, if you stick to the web-safe color palette, you will avoid causing
enlarged 200%.)
dithering and its resulting illegibility and aesthetic problems on 8-bit sys-
tems. As explained in Chapter 2, the practice will not help those with 16-
bit systems, but nothing can save those folks except a graphics card
upgrade in their future.
13 0732 CH09 4/24/01 11:22 AM Page 215
Death of the Web-Safe Color Palette?

Creative people complain about everything. Web designers certainly com-
plain about being limited to 216 web-safe colors, but to us this is like grip-
ing about the nip in the air while enjoying the scenic beauty of rustic New
England. You want fall foliage, so put on an extra sweater.
Lulled by the music of these constant complaints, pundits perennially pro-

claim the death of the web-safe color palette, usually on the grounds that
16-bit systems enjoy a major market share. That 16-bit systems are widely
used is undeniable: They are installed in 46% of PCs as of this writing. That
the web-safe color palette is therefore dead is wishful thinking.
The web-safe color palette cannot die as long as it continues to solve prob-
lems for millions of web users. It does not solve every problem, but neither
does penicillin, and nobody talks about the death of penicillin. We bring
this up now because you will hear about it at the office and read about it
in web design newsletters, mailing lists, and bulletin boards.
Who spreads these obituaries? Sometimes it’s information architects and

interface developers who conduct meaningful research but draw debatable
conclusions from their data. The Webmonkey article, “Death of the Web
Safe Color Palette?" (http://hotwired.lycos.com/webmonkey/00/37/
index2a.html), proves beyond all doubt that 16-bit systems are hopelessly
inadequate and invariably reveal the rabbits hiding in a web magician’s hat.
But the article nihilistically concludes that all color palettes and traditional
methods are meaningless in the chaos of the Web; whereas we judge sim-
ply that 16-bit users are hosed until they upgrade. Not long ago, 16-bit
color was considered luxurious; cheap graphics cards changed the market,
and the next generation of cheap 24-bit cards will change it again.
Few discussions of the topic have been as carefully researched as Web-

monkey’s. The death of the web-safe color palette is generally announced
by the same people who tell us that bandwidth no longer matters because
“everybody” will “soon” enjoy high-speed access. These folks often go on to
proclaim that presently every site will be pumping out full-screen video
productions to rival Hollywood blockbusters.
13 0732 CH09 4/24/01 11:22 AM Page 216
A moment’s analysis will tell you that many people around the world are
not online yet. That those who are online are mainly limited to slow con-
nections over untrustworthy phone lines. That even in the major urban
areas of industrialized nations, high-speed access is often hard to come by
and frequently comes at a premium many cannot afford—or are not will-
ing—to pay. That major Hollywood productions cost millions and can make
a profit (when they do make a profit) only by charging admission. That web-
sites generally do not charge admission, and web clients generally do not
have millions of production dollars at their disposal. And finally, that most
people do not seek big-budget entertainment from the Web. They seek
information, services, and communities—all of which the Web can deliver
with a minimum expenditure of bandwidth.
In other words, much of what you hear about how the Web works and
where it is going is bunk—including, we think, the death of the web-safe
color palette. Ask us again in a year or two when (hopefully) most PCs come
standard with 24-bit color or higher.
A Hex on Both Your Houses

Reared on RGB and CMYK, many designers find the Web’s hexadecimal
color nomenclature strange, at first. But the predictability of recurring
hexadecimal pairs (00, 33, 66, 99, cc, ff) makes it easy to tell if you are
using web-safe colors or not. It also makes it easy to specify web-safe
background colors and text colors in HTML and CSS.
You will find, after you work with these colors, that it is possible to create
pleasing combinations with them, and you will develop your own tech-
niques for doing so.
In this quest, you will be greatly aided by Photoshop’s own tool set and by
the VisiBone color palette included in Photoshop 5.5 and higher and avail-
able free online. The VisiBone palette is a superb tool for establishing visual
relationships between web-safe colors. And, as you already know, visual
relationships are the key to creating pleasing and effective color schemes.
13 0732 CH09 4/24/01 11:22 AM Page 217
Color relationships are essential to branding, can support navigational

structures, and may greatly enhance a site’s aesthetic appeal at a minimum
expenditure of bandwidth. Fill an entire page with a CSS background color,
devise complementary link and text colors, and you begin to have the rudi-
ments of an attractive design using just a few kilobytes of bandwidth.
Was Blind, but Now I See

It should be noted that a small percentage of women and a larger per-
centage of men suffer from various forms of color blindness. Designs that
rely exclusively on color to convey essential information and relationships
could therefore be inaccessible to some viewers. So, while taking color
extremely seriously, you must also test your designs for accessibility—
ideally by running them past test subjects who manifest different forms of
color blindness. If you can’t do it that way, use the Color Blindness Simu-
lator at www.vischeck.com. Viewing your web layouts in grayscale mode is
a nice gesture but not a truly accurate means of testing how they will
appear to, say, a person with red/green color deficit (deuteranopia).
From Theory to Practice

The following three exercises introduce you to the effects of dithering,
describe how to set up the Photoshop Color Picker so that your color
choices are always web-safe, and explain how to locate and install the Vis-
iBone color palette. You will notice that we begin each exercise by cau-
tioning you to set your monitor to 24- or 32-bit mode before launching
Photoshop. If you accidentally launch Photoshop while in 16-bit mode, all
your colors will shift, and the images you design for the Web will always
be mismatched from their backgrounds.
13 0732 CH09 4/24/01 11:22 AM Page 218
Exercise 1: In a Dither
Be certain your monitor is set to 24- or 32-bit mode. Launch Photoshop 5.5
or higher.
Open a new, blank document (600 x 400 pixels) and paint in it randomly, using
the Paintbrush and Airbrush tools.
Also be sure to use the Type tool to set some large type in a variety of colors.
Stop when you are satisfied.

In the Mac Finder or Windows desktop, switch your monitor to 256 colors.
Look at the image you’ve created. Those ugly dots are dithering, and that’s
what millions of viewers will see if you do not learn to incorporate the web-
safe color palette into your work.
Close the image without saving it, quit Photoshop, and restore your monitor
to its normal color settings (millions of colors).
Exercise 2: You Sure Can Pick 'em
or higher.
Open a new, blank document (600 x 400 pixels).

Open Photoshop’s Color Picker (see Figure 9.3). Note the Only Web Colors
checkbox and check it.
Figure 9.3
The Photoshop Color
Picker provides RGB, HSB,
Lab, CMYK, and hexadeci-
mal readouts for any color
you choose. The familiarity
of RGB and CMYK will
help acclimate you to
hexadecimal nomencla-
ture. Click the checkbox
that reads Only Web
Colors, and your choices
will always be web-safe.
13 0732 CH09 4/24/01 11:22 AM Page 219
Watch your universe of color options shrink down to 216 choices. On the plus
side, the various graphic dialogs help you see the relationships between web-
safe colors.
Close the Color Picker dialog.
From now on, Photoshop’s Color Picker will always be web-safe. You can also
use Photoshop’s Color Picker to shift a near-web-safe color so that it is fully
web-safe.
Photoshop’s web-safe Color Picker is a vast improvement over what the pro-
gram used to offer in the way of support (namely, nothing).
Now that our Color Picker is web-safe, let’s do the same for our color palette
dialog. Jeepers, but we are moving along quickly here.
Exercise 3: Rolling the ‘Bones
or higher.
Open a new, blank document (600 x 400 pixels).

Refer to the Colors dialog. Note that there is a web-safe palette included in
Photoshop. Note that the color combinations are not especially intuitive and
have no meaningful relationship with the color wheel or other color theory
models.
Let’s fix that.
In the Swatches dialog box, choose Replace Swatches. A dialog box opens,
allowing you to navigate to a new palette located on your hard drive. Steer
your way to the VisiBone color palette (VisiBone1.aco), which is most likely
located in Adobe Photoshop 5.5, Goodies, Color Swatches.
Handsome, isn’t it? (See Figure 9.4.)
13 0732 CH09 4/24/01 11:22 AM Page 220
Figure 9.4
The VisiBone color palette,
located in the Color
Swatches folder within
the Goodies folder in
Photoshop 5.5 and higher,
makes it easy to choose
harmonious or contrasting
colors from within the
web-safe palette. Don’t
leave home without it.
This is still the web-safe color palette. But unlike Photoshop’s built-in, default
version, the VisiBone palette offers a meaningful arrangement built around
the color wheel model we all learned about in school (unless we spent our time
in school doodling and learned almost nothing except how to draw guitars in
the margins of our textbooks). Colors move in a circle across the spectrum,
and related colors are geometrically aligned with respect to one another.
The VisiBone color palette not only helps you choose web-safe colors, it helps
you choose web-safe colors that relate to one another in a meaningful way,
man. Harmonious and contrasting color relationships are easy to see and thus
easier to create.
In other words, the VisiBone palette helps you start doing beautiful work
within the limitations of the Color Cube. For instance, it helps you quickly find
a web-safe approximation of a client’s logo color and begin experimenting
with complementary and contrasting web-safe colors for your layouts. (It goes
without saying that if original logo development is part of the project, you will
design the logo using web-safe colors. It also goes without saying that your
client might want you to design a logo that matches the color of their
new Beetle or their favorite coffee mug, but that is where tact and client
education come in.)
13 0732 CH09 4/24/01 11:22 AM Page 221
Save the VisiBone swatch so it is always available when you work in

Photoshop.
Now pat yourself on the back. Many of your peers have no idea that this spe-
cial swatch exists, that it comes bundled with Photoshop, and that it can
greatly ease the creation of meaningful and attractive color schemes for the
Web. You are ahead of the game.
If you’re stuck using an older version of Photoshop or an alternative image

editor, you can download the VisiBone palette free of charge at
www.visibone.com. While there, help yourself to additional VisiBone
palettes for other software programs you or your teammates use, includ-
ing Adobe Illustrator and ImageReady, Macromedia Fireworks, Bare Bones
BBEdit, Jasc Paint Shop Pro, Allaire HomeSite, MetaCreations Painter, or the
GIMP (an image editor for Linux). You need it; they’ve got it.
For additional wisdom on the Color Cube, see Lynda Weinman’s site at
www.lynda.com and David Siegel’s at www.killersites.com. You also might
want to buy Weinman’s Designing Web Graphics and Coloring Web Graph-
ics, both of which are available from New Riders Press, and are pretty much
the standard industry texts. They are full of practical examples and offer
stimulating and innovative ideas from the earliest days of web design.
Another standard industry text, David Siegel’s Creating Killer Websites, is

also available from New Riders and also provides extensive information on
the subjects we cover in this chapter. It’s a beautifully written book full of
great ideas, but it is also a book of its time (1996), and many of the prac-
tices it preaches would now be considered harmful to the development of
a semantic Web based on W3C Recommendations. We own and cherish this
book, which was greatly influential in our development, and we recom-
mend it as long as you know which of its visual techniques to shun. (If
you’re unsure, wait for the book’s third Edition…we hear it’s coming soon.)
FORMAT THIS: GIFS, JPEGS, AND SUCH

Raster images come in at least as many formats as there are software pro-
grams and operating systems. On the Web, however, we tend to use two
formats almost exclusively: GIF and JPEG. (As explained previously, ani-
mated GIFs are a special instance of the GIF format.)
13 0732 CH09 4/24/01 11:22 AM Page 222
222 HOW: Visual Tools: Format This
PNG is yet another web format, one that has been little supported in the
past. Some newer browsers have begun to support PNG, though it is still
far from ubiquitous. We will discuss it after thoroughly examining the GIF
and JPEG formats—how they work, which types of images they deliver best,
and how you can evolve strong stylistic concepts by understanding their
limitations.
GIF
The Graphics Interchange Format (GIF) is older than the Web. In fact it is
older than some web designers. GIF was developed in the 1980s by Com-
puServe, and you’ll often hear old-timers speak of “CompuServe GIFs.”
You’ll also hear them talk about walking 12 miles to a one-room school-
house.
The Compuserve folks pronounced the word as if it were the name of the
peanut butter (“Jiff”) and because they were the inventors, that is the cor-
rect pronunciation. Millions of people pronounce GIF with a hard “G,” how-
ever, so you might as well be a sniveling conformist and spend the rest of
your career mispronouncing GIF while secretly suffering great guilt over it.
GIFs are usually seen with a .GIF file extension, as in payme.GIF or
payme.gif.
The GIF format renders in 8-bit color or lower, at your discretion. Two-color
GIFs are not uncommon. GIF permits you to achieve crude transparency
effects by marking one of your 216 (or fewer) colors as “transparent.” How-
ever, you must take care to anti-alias the foreground image against the
transparent color, lest mismatched halos surround your graphics. Fortu-
nately, GIF renders specific colors exactly, so it is an easy matter to match
web page backgrounds to image backgrounds. The only caveat there is the
previously mentioned heartbreak of 16-bit systems.
Above all, GIF enables you to save bandwidth without sacrificing quality. It
employs the Unisys-patented Lempel Ziv Welch (LZW) algorithm
(www.dogma.net/markn/articles/lzw/lzw.htm) to efficiently compress solid
color areas while preserving crisp detail. Though the format necessarily dis-
cards colors—for instance, when rendering a 24-bit image as a 16-color
GIF—it does not blur or eliminate significant image details. For this reason,
the GIF algorithm produces what is known as lossless compression.
13 0732 CH09 4/24/01 11:22 AM Page 223
Loves logos, typography, and long walks in the woods

This combination of crisp detail and efficient compression makes GIF the
format of choice for line art including typography, logos, and illustrations.
As mentioned earlier, the GIF format can also be used to create animated
images. When combined with JavaScript rollovers, animated GIFs can lend
life and dynamism to a website. They can also create nausea and ennui.
With animation and rollovers, as with Tabasco, a little goes a long way.
Animated GIFs have been supported in all graphical web browsers since
Netscape 2.0 (1995), and nonanimated GIFs have been supported in graph-
ical web browsers since before time began. For now we will continue to
discuss the merits and uses of static (nonanimated) GIFs.
In spite of the fact that GIFs are found on millions of sites, the GIF format
is not a W3C-recommended web standard. That’s because GIF gets its
power from a patented algorithm. Unisys, the patent holder, is entitled to
charge royalties on any software that employs the LZW algorithm—in other
words, any software that can read or write GIFs. The revelation of Unisys’
right to charge a “GIF tax” spread panic among early web designers when
it became widely known only after the entire Web seemed to be built with
GIF images. It also led to the development of PNG, a GIF-like format with
more advanced features and a nonproprietary compression algorithm.
GIF “royalties” do not work in the way that, say, photo rights work. You do
not pay a fee each time you create a GIF image. Instead, software compa-
nies such as Adobe, Macromedia, and Corel render these tributes to Cae-
sar. You pay your share one time only, and it is hidden in the purchase price
of Photoshop, Fireworks, or any other software program that exports to the
GIF format.
GIFs are not the format of choice for photography, paintings, and other
subtly modulated images because they lack sufficient colors to reproduce
these types of images and because the nuances in those images do not lend
themselves to LZW compression. Photographic images tend to render bet-
ter in the JPEG format (or PNG), and we’ll get to those formats soon
enough.
13 0732 CH09 4/24/01 11:22 AM Page 224
GIFs in Photoshop
In Photoshop, you can choose whether to save your image as a standard or
interlaced GIF. The standard format is like a reader, taking in one letter after
another, one word after another, one sentence after another. Standard GIFs
store and display the bytes comprising an image’s pixels in their order of
appearance: The first pixel in is the first pixel out. Thus, standard GIFs scroll
onto the viewer’s screen pixel by pixel and line by line.
The interlaced format is like a nervous reader who keeps skipping ahead—
from paragraph one to paragraph five, then back to paragraph one. Inter-
laced GIFs load in a parallel rather than linear sequence, allowing the total
image to be rendered more quickly and then with greater detail as addi-
tional pixels are downloaded. This allows viewers to get a sense of the
image before it has finished downloading.
Under the right conditions, interlaced GIFs might thus appear to load
faster—and so may your site. The appearance is deceptive given that inter-
laced GIFs are often a few bytes larger than standard GIFs and therefore
take a fractionally longer time to fully download. Moreover, the slight
benefits of interlaced GIFs often evaporate when other conditions are
factored in.
For one thing, the effectiveness of progressive GIFs depends on the viewer’s
access speed. With a super-fast connection, images load so quickly that
any progressive rendering benefits are lost. The format was something of a
godsend not so long ago, when most web users were limited to 14.4
modems. Today, few are stuck with such abysmal speeds.
The effectiveness of progressive GIFs also depends on the browser. Some

browsers do not show anything at all until all images are fully loaded; in
those browsers, the progressive aspects of the image are entirely wasted.
If anything, in such browsers, progressive GIFs delay the page by adding a
few bytes to the overall download time.
Some browsers, such as Internet Explorer, give users a choice. Users may
view each image as it downloads (best with slow connections), or they may
choose to wait for the entire page to download and assemble itself in mem-
ory before appearing full-blown on the screen (best with fast connections).
Users choose a viewing method in the Explorer Preferences dialog box. You
have no way of knowing or controlling these user preferences.
13 0732 CH09 4/24/01 11:22 AM Page 225
Beginning web designers often ask if they can control the loading order of
images on a web page. Given what has just been explained, the answer is
obviously “no,” because web users can choose (or their browsers may force
them) to wait for the entire page to load. Beyond that, HTML has no means
of controlling the loading order of images. And even if it did support such
nuances, the unpredictability of HTTP calls (explained in Chapter 2) means
that one image might halt in mid-download, not even appearing until
another, called much later, has already popped into place. The more images
per page, the greater the randomness of load order. View a busy thumbnail
image gallery sometime to see this in action, assuming your browser allows
you to watch images download one by one.
Avoid progressive GIFs when creating an image to be used as a background.

Backgrounds do not appear until they have fully downloaded, so any “pro-
gressive” effects will be lost. Moreover, progressive GIF backgrounds can
crash some older browsers.
Progressive GIFs also can be hazardous to animations because each suc-

ceeding frame of a progressive animated GIF will appear blurry, thus
defeating the effort to create smooth motion effects.
They’re not great for JavaScript rollovers, either. You can offset the harm-
ful, blurred quality of progressive GIFs in rollovers by preloading the
images, a technique explained in the Chapter 11, “The Joy of JavaScript.”
When preloaded via JavaScript, images download and are stored in the
viewer’s cache even though they do not appear on the web page until trig-
gered by some action on the viewer’s part (typically, moving the mouse over
an image to which rollover effects have been applied). Any sane web
designer who creates rollovers starts by preloading the alternate (replace-
ment) images. But if the images are going to be preloaded anyway, there’s
no sense in having them render progressively because the user will never
see them until they have fully downloaded and cached.
One last tip while we’re in this area. Given that text loads instantly and
images take time (see Chapter 2), designs that use HTML text above the
fold will appear to load more quickly than those that bury their text fur-
ther down on the screen. A web user waiting for images is a web user with
nothing to do (except, perhaps, hit the Back button). A web user reading
13 0732 CH09 4/24/01 11:22 AM Page 226
text has less anxiety about the fact that some images may not have fin-
ished downloading. With sufficiently engaging text, the user will feel that
the site is responsive. Keep this in mind when designing sites that require
a great many images.
JPEG, the Other White Meat

The Joint Photographic Experts Group (JPEG) format should be familiar to
you from stock photo houses, digital cameras, and the Photoshop tutorial
itself. Usually seen with a .jpg file extension (as in landscape.jpg), JPEG sup-
ports 24-bit color and preserves the subtle hue and brightness variations
found in photographs and other continuous-tone images. JPEG is therefore
usually the format of choice when creating photographic images for the
Web. Like GIF, JPEG is widely supported in visual web browsers.
Just as MP3 music files toss away audio harmonics to achieve compact file
sizes, JPEG’s compression works by selectively discarding bits of image
data. Because a loss of quality is involved, JPEG compression is referred to
as lossy compression. “Lossy” is an annoying word that looks wrong, but we
appear to be stuck with it. In theory, the material discarded by the JPEG
optimization process is data that is nearly invisible to the human eye (just
as audio data discarded by the MP3 format is supposed to go practically
undetected by your ears, though we’ve never met a music fan who could
not hear the difference). The greater the JPEG compression, however, the
more visible the “missing data” becomes. At extremely high compression
ratios, JPEG images can display funky artifacts (see Figures 9.5 and 9.6).
Although JPEG is generally preferred for photographic images, when sharp

detail is important, GIF is the better choice. JPEG tends to soften images as
it compresses them. Particularly when you are working with typography,
the softness of JPEG images can ruin the effect of a web graphic. Naturally,
there is a workaround, as explained in the “Combining Sharp and Blurry”
section later in this chapter.
Unlike GIF, the JPEG format does not retain specific web-safe (or other) col-
ors. It promises you a rose garden, but the rose might be umber. In a sil-
houetted portrait where the edges of the image must match the
background of the web page, you would therefore use GIF, not JPEG.
13 0732 CH09 4/24/01 11:22 AM Page 227
Figure 9.5
At moderate JPEG
compression levels,
image details are clear,
but file size is high.
Figure 9.6
At high JPEG compression
levels, file size is low
(minimizing bandwidth)
but so is the quality. Each
JPEG optimization is an
exercise in balancing file
size versus quality of
detail.
13 0732 CH09 4/24/01 11:22 AM Page 228
Photoshop’s Save For Web function provides a small, Matte Color dialog
box that purports to save an exact background color of your choice, even
in the JPEG format. (Skip ahead to Figure 9.7, if you must. The Matte Color
dialog appears at mid-right.)
Photoshop does all it can to fulfill this promise, but the JPEG format really
is not built to handle specific colors like this. To viewers with 24-bit and
higher systems, the background color will appear to match. For 16-bit and
lower users, the mismatch may be clearly visible. So stick with GIF when
you absolutely, positively, have to deliver a specific web-safe (or other)
color.
In Photoshop, you can choose whether to save your JPEG as a baseline

(standard) JPEG or as a progressive JPEG. Progressive JPEGs display a low-
resolution version of the image almost immediately and then gradually
come into crisper focus.
As with progressive GIFs, under the right circumstances, progressive JPEGs

can create the illusion that the site is loading faster. As previously dis-
cussed, this varies depending on the viewer’s access speed, browser func-
tionality, browser preferences, and the caprices of HTTP. And as in the
discussion of GIFs above, when intended as background images, progres-
sive JPEGs are a no-no unless you want some of your visitors to crash-
crash.
Optimizing GIFs and JPEGs

When we export images such as GIFs and JPEGs, we choose the format
most appropriate to the type of image we’re dealing with and then opti-
mize it to create the best appearance possible, while using the least
amount of bandwidth and computing resources.
In addition to optimizing (reducing file sizes), the exporting process allows

us to further exert control over the color of our GIF images.
Photoshop 3, 4, and 5 offered early web designers very little in the way
of optimization and color controls. As a result, a number of inexpensive,
third-party, shareware plug-in products specifically tailored to the needs
of web designers sprang up in the mid-1990s, most notably Boxtop
13 0732 CH09 4/24/01 11:22 AM Page 229
Software’s PhotoGIF, ImageVice, and ProJPEG (all are available at

www.boxtop-software.com). These products were dandy (still are), but they
did not come as standard equipment (still don’t). Arguably, they do a bet-
ter job than Photoshop at handling some tasks.
Fortunately, Photoshop 5.5 and higher, together with ImageReady, offers a

number of tools to help web designers create the best-looking image while
using the least amount of bandwidth. Photoshop’s Save For Web command
(found in File, Save For Web) enables web designers to preview the effects
of various compression settings on their images and then execute those
settings and save the resulting web-ready images (see Figure 9.7).
The Save For Web dialog is powerfully compelling in the breadth and sub-
tlety of its tools. You can preview GIF versions using as few as two colors,
as many as 256, or anything in between. You can use Adaptive, Selective,
Perceptual, or web-safe color, with or without dithering, transparency, or
interlacing (the “progressive” setting). You can skew images closer to or
further from the web-safe palette as you desire. You also can name and
save custom settings for later application to similar images.
Figure 9.7
Photoshop’s Save For
Web dialog in action. In
this “four-up” view, the
original image appears at
the upper left for easy
comparison with various
optimization schemes of
your choosing.
13 0732 CH09 4/24/01 11:22 AM Page 230
Work on one optimization setting at a time or view three at once—and

compare them with the original to check for image degradation and color
shifting. Get an instant readout of the effect your decisions will have on
file size and downloading speed. Enlarge images to check fine details. Lock
selected colors before trying a new set. Shift one color at a time to its clos-
est web-safe equivalent. We feel like press agents. We feel giddy. We love
this dialog box. You will too.
Images in Save For Web mode also may be previewed at various JPEG set-
tings, both baseline and progressive, and again the tools are remarkably
powerful.
In general, the fewer the colors used in a GIF, the better it compresses. This
is not because the color palettes themselves eat bandwidth; rather it is
because of the way LZW compression works. More on that in a moment in
the “Expanding on Compression” section that is coming up next.
Dithering images produces more photographic-like effects at the cost of

slightly higher file sizes; images without dithering are smaller. We find that
typographic GIFs are often cleaner and more legible when saved without
dithering. Your mileage may vary. You can create either type of image (and
preview the results) in Photoshop’s Save For Web dialog box.
After you decide which optimization scheme works best for a given image,
the image can be saved in that format. Your chosen settings may be
retained indefinitely, and can even be applied (as a droplet) to an entire
folder of images.
Photoshop lets you name and store as many of these settings as you like.
If a series of images you’ve created for Acme Widgets happens to work well
in 12 colors with no dithering at 60% web-safe, you can name that set-
ting 12color_nodither (or acme_widgets or 60websafe or donaldduck if
you prefer). You can then save it forever—or at least until your backup
media deteriorates and what’s left of your hair is white and listless. By then
we’ll all be living on Mars while our clones do the work, anyway.
Alternately, you can use the ImageReady module to satisfy your wanton
image compression and formatting needs. But Photoshop’s Save For Web
is just as effective, and the true power of ImageReady comes later in the
process (and this chapter).
13 0732 CH09 4/24/01 11:22 AM Page 231
Expanding on Compression
As explained previously, we compress images to minimize wasted band-
width and speed the arrival of the web page. We’ve shown you how Pho-
toshop optimizes images when preparing them for the Web, so you know
all you need to know to handle the basics. The following are some extra
tips.
Make your JPEGS smaller

You can make your JPEGs even smaller in file size (and reduce the appear-
ance of JPEG artifacts) by blurring them slightly before compressing them.
Not all areas of all images react well to blurring, but it’s surprising what
you can achieve by blurring, say, a distant sky and sunset while preserving
the sharpness of a human subject in the foreground.
This kind of work requires selecting the parts of the image you want to blur,
feathering the edge of the selection slightly, and masking out the parts of
the image where sharper focus is important. As we said in the beginning,
we assume you already know how to do these things in Photoshop.
If you prefer, you can apply subtle (or not-so-subtle) blurring effects to
your entire image in the Save For Web dialog box, but we generally find
this method too coarse. Blurring, say, an entire portrait makes the subject
look drunk—or the viewer feel that way. Selectively and subtly blurring
large areas of undifferentiated skin tones, while preserving the sharpness
of eyes, brows, hair, and lips, will usually be much more effective. And that
kind of work you do in the main Photoshop window before entering the
Save For Web dialog.
Combining sharp and blurry

Subtle problems can arise when choosing the appropriate image format.
Say you’ve designed a header graphic that includes both photography (a
shot of the corporate board of directors) and typography (a superimposed
headline in Meta or Helvetica Neu Condensed Black). The headline requires
sharp focus and crisp handling—thus it begs to be a GIF. The photograph
wants to be a JPEG. What’s a mother to do?
13 0732 CH09 4/24/01 11:22 AM Page 232
Usually, what you do is give greater weight to the need for crisp typogra-
phy and export the entire image as a GIF, accepting that the photographic
imagery will not render as well as it would have in a JPEG (see Figures 9.8,
9.9, and 9.10).
Figure 9.8
The background image,
a layered photomontage,
wants to be saved as a
JPEG because JPEG would
best reproduce its subtly
modulating hue and
brightness variations.
But…
Figure 9.9
…the typography insists
on being saved as a GIF
because only the GIF for-
mat will reproduce the
crisp, clear lines of type.
A JPEG would soften the
headline and render the
small type as an illegible
blur. So…
Figure 9.10
…the image is saved as a
GIF because type takes
precedence over photo-
graphic nuances. The
image could also have
been saved as a PNG. But
the PNG would have been
far larger and not enough
browsers fully support the
PNG format yet.
13 0732 CH09 4/24/01 11:22 AM Page 233
Alternately, you could export the human subjects as a JPEG, export the
typography as a transparent GIF, and superimpose the GIF over the JPEG
using any number of web sleights of hand. For instance, you could employ
CSS absolute positioning to layer a crisp, transparent typographic GIF on
top of a soft photographic JPEG. (This would not work in Netscape 3 or IE3
and might destabilize Netscape 4. Thankfully, these browsers are finally
limping away from the playing field, although not as fast as we’d like.)
Depending on the layout, you alternatively could use the old, nonstandard
<bgimage> attribute of the HTML table cell <td> tag to position a photo-
graphic image in the background of a table cell and then place a type GIF
in the foreground. The type GIF would have to be the same size as the back-
ground image and would require GIF transparency to allow the background
to peek through. The size of the GIF would be marked up in the table cell
attributes to ensure that the cell was the correct size. Though this tech-
nique works well in almost every graphic browser since the svelte boyhood
of Fred Flintstone, it is a lot of silly (and nonstandard) markup—and it’s
probably not worth the bandwidth.
Or you could do what Magdalena Donnea did on the front page of her
award-winning personal site, “Water,” at www.kia.net/water/. (Use View
Source to see exactly what Magdalena did.)
As we said, most of the time, you’ll use the GIF format to ensure that your
text is legible. You also might consider rethinking the entire design idea in
favor of one that is more in keeping with the limitations of the Web (see
Figure 9.11).
Figure 9.11
The ever-popular “striped”
effect that dominated the
web in the late 1990s had
its roots in a technique to
minimize bandwidth by
making the most of the
GIF compression algo-
rithm’s preference for
straight horizontal lines.
13 0732 CH09 4/24/01 11:22 AM Page 234
234 HOW: Visual Tools: Compression Breeds Style
COMPRESSION BREEDS STYLE: THINKING

ABOUT THE MEDIUM
The GIF format not only compresses by removing millions of colors, it also
employs the LZW algorithm to keep track of those colors and further reduce
file sizes. A clever web designer can create large images that use little
bandwidth by designing with LZW compression in mind. To understand how
that is possible, we must take a closer look at how LZW compression actu-
ally works.
Onscreen images are like diners inside a burger joint. A mentally challenged
waiter says, “The first gentleman at Table One would like a cheeseburger.
The second gentleman at Table One would like a cheeseburger. The third
gentleman at Table One would like a cheeseburger. The fourth gentleman
at Table One would like a cheeseburger.”
This is how a noncompressed image works. The computer looks up the color
of a pixel and then displays it. It looks up the color of the adjoining pixel
and displays that—and so on and so on for every pixel on the screen.
A smart waiter says, “Four cheeseburgers,” which is how LZW compression

works.
LZW compression looks at an image line by line and says, “Row #1 is all red
pixels” (assuming that Row #1 actually is all red pixels). Obviously, the
greater the number of pixel rows that are identical to each other, the bet-
ter the compression engine works (for example, four tables of four cheese-
burgers). Thus, horizontal and vertical elements compress better than
diagonal elements because with horizontal or vertical elements, more rows
of pixels can be exactly the same as each other.
Without getting too technical, horizontal lines tend to compress even bet-
ter than vertical ones because LZW compression “reads” images left to right
and line by line, the same way you’re reading this book. If every pixel in a
given line is the same color, that line compresses better, and therefore so
does the GIF (there’s more to it than that, of course). Ten lines containing
13 0732 CH09 4/24/01 11:22 AM Page 235
all the same color compress better still. Basically, GIF compression likes
large areas of flat color, whether they are confined to a single line or bleed
down across several. The main point is that an image containing one or
more lines of identically colored pixels will compress much better than the
average image whose colors are arrayed at random.
In 1995, when 14.4 modems prevailed, some clever web designers began
masking every other horizontal line in an image to maximize LZW com-
pression and minimize bandwidth. This technique of masking images
with evenly spaced horizontal lines is known as CRLI compression
(www.infohiway.com/faster/crli.html).
What started out as a bandwidth-oriented tool had become a stylistic

design fetish by the late 1990s, as newcomers to the field fell in love with
the CRLI “look” without understanding its utilitarian purpose as a tool of
bandwidth compression. To these designers, stripes were “webby,” and
“webby” was cool. As the Web exploded into public consciousness, con-
sumers and ad agencies seemed to agree with this link between “Web” and
“cool.” The ever-popular striped effect was soon seen not only all over the
Web, but also in print and television.
Among many ironies, some web designers exported these striped images in
the JPEG format, where, far from saving bandwidth, the technique actually
wasted it. They knew not what they were doing. CRLI compression is a GIF
thing, baby.
The strengths and limitations of LZW compression are equally profound. For
instance, because LZW prefers straight horizontal and vertical lines to all
others, Roman type tends to reproduce better than oblique. Roman type is
also better at hiding its anti-aliasing artifacts at screen resolutions—
another reason it works better onscreen than oblique does.
Considering these limitations of the medium may lead you to set your
headlines in Roman type more often than oblique. Of course, Roman type
is far more frequently used than oblique to begin with, so this situation is
hardly tragic. But you should be aware of it. Oblique type can certainly be
used for headlines—we do it all the time—but it never reproduces as well
as upright type.
13 0732 CH09 4/24/01 11:22 AM Page 236
236 HOW: Visual Tools: Compression Breeds Style
You will run into the same difficulty with lines at almost every angle. The
45-degree angle is the exception: It works perfectly with LZW, like a diag-
onal in a game of tic-tac-toe. As you might expect, 45-degree angles came
into vogue around 1999 because they reproduce well on the Web, and
within six months they were popping up in print and TV as a meaningless
design fetish after everyone had tired of the striped effect. And as you
might also expect, many web designers employed 45-degree angles in
JPEGs, then saved the JPEGs at the highest possible quality settings to pre-
serve the crispness of their lines. The result: wasted bandwidth.
PNG
The PNG format was developed in hopes of establishing it as an open
standard for graphics on the Web—which it now is (see www.w3.org/
Graphics/PNG/). But while PNG was slowly being developed, working web
designers had to create websites, and all browsers supported GIFs. In effect,
then, GIF is a long-standing, unofficial defacto standard based on a pro-
prietary compression algorithm, while PNG is a nonproprietary, officially
sanctioned standard that is not as well supported as it ought to be.
There are two forms of PNG. PNG-8 is an 8-bit format (like GIF). PNG-24
offers 24-bit color (like JPEG), yet its sharpness and quality put JPEG to
shame. To create PNG images for the Web, simply choose PNG-24 or PNG-
8, 128 Dithered in Photoshop’s Save For Web dialog box or in ImageReady.
PNG is still not natively supported in enough web browsers, and though
support is growing, PNG is unlikely to supplant GIF or JPEG any time soon.
For one thing, PNG, while high in quality, is often high in bandwidth as well.
For another, while PNG stays crisp in milk (like GIFs do), the PNG format
does not support animation. GIFs are therefore seen as more versatile by
those who even bother to lift their heads out of their cubicles and think
about these issues.
To see why PNG can be cool indeed, if your browser can handle it, visit the
Audio site at www.panic.com/audion/faces.php, click any thumbnail, and a
PNG image will pop up on the screen. Drag the image from place to place
on the page at your pleasure. You can even drag it off screen (as shown in
Figure 9.12).
13 0732 CH09 4/24/01 11:22 AM Page 237
Figure 9.12
PNG a ding-ding. On the
Audion site, you can bask
in the glories of the PNG
format—glories that
include true alpha channel
transparency, rich color,
and crisp detail. (But only
if you’re packing the right
browser.)
Notice that the PNG format offers true alpha channel transparency—it
matches any background you drag it over. No more halo effects caused by
mismatched anti-aliasing, no more ring around the collar. Notice too that
PNG offers crisp imagery as well as rich color.
Notice that the page only works in IE5 for the Macintosh. Bummer. Even-
tually all browsers will support PNG natively.
ANIMATED GIFS
Animated GIFs are nothing more than a series of frames (or individual GIFs)
that have been joined together to create the illusion of motion. They can
loop endlessly or play once and then stop. We could include a screenshot
here, but what’s the point? If you haven’t seen animated GIFs, you’ve never
used the Web. (Hint: look at the ad banners that clutter most commercial
content sites—web animation in a nutshell.)
Although the GIF format supported the embedding of multiple images in

the late 1980s, it was not until 1995 or so that Netscape figured out how
to hack the format’s multi-image capability to create flip-book-style ani-
mation. (Basically, Netscape did this by appropriating a Comments field and
some unused but reserved bits in the GIF89A file format.)
13 0732 CH09 4/24/01 11:22 AM Page 238
238 HOW: Visual Tools: Creating Animations in ImageReady
Back in the day, web designers used free shareware tools to create ani-
mated GIFs, after first preparing each individual image, saving it as a GIF,
and then running all resulting GIFs through DeBabelizer, a cumbersome
color management tool that ensured that the colors would match between
frames. (Nothing ruins the illusion of motion faster than an unexplained
color shift between one frame and the next.)
Today all that work is merely a memory because Photoshop comes with
ImageReady, and ImageReady makes it easy to create, optimize, and save
GIF animations.
Animation for its own sake is charmless, abrasive, and amateurish. Good
web designers use animation as they use everything else: with taste and
skill in support of a concept and brand image. The creators of www.k10k.net
employ animated GIFs well. The animations are revealed when rolling over
the miniature content header graphics.
Care should be taken to avoid wasting bandwidth when creating animated

GIFs. If one image uses x bytes, then ten images theoretically use 10x bytes,
and your web page might bloat as a result. Fortunately, web designers can
trim excess fat from their animations by telling the software to animate
only the parts that change, rather than redrawing each frame in its entirety.
This process is explained in the next sections. Web designers also can opti-
mize their animations by leaving out inessential in-between frames, by
keeping their images small (50 x 50 is better than 100 x 100), and by cre-
ating graphics that can be rendered in as few colors as possible.
CREATING ANIMATIONS IN IMAGEREADY

Adobe ImageReady simplifies the process of creating animated GIFs by
allowing web designers to use Photoshop’s layers as a series of frames and
enabling them to manually change the location of elements from one
frame to the next.
For instance, if you wish to animate an arrow, you can draw the arrow on
one layer in Photoshop then jump to ImageReady and open the animation
palette. Create a new frame and drag the arrow manually to the left or
13 0732 CH09 4/24/01 11:22 AM Page 239
right. Create a third frame and drag the arrow again. ImageReady “remem-
bers” the location of each arrow and will render an animation as a result
of these manual movements.
ImageReady can also generate tweens automatically. Start with an arrow

on the left. Create a new frame. Drag the arrow to the right. Choose the
Tween command and instruct ImageReady to tween between the first and
second frames. ImageReady generates a smooth flow of images. You can
then use the Optimize palette to ensure color consistency from the first
frame to the last. Keep in mind that the more you tween, the smoother the
motion but the larger the overall file size.
We could blab on about this, but the Photoshop owner’s manual does a
great job of explaining everything. The way we see it, if you own Photo-
shop, read the manual. If you don’t own it, there’s no sense in reading about
it here and probably not much sense in planning a web design career. (Gosh,
that sounds like a product endorsement.)
TYPOGRAPHY
A designer’s interest in typography usually borders on obsession. On the
Web, you’ll get plenty of opportunities to indulge your fetish. As part of
establishing the look and feel of a site, the web designer is responsible for
all of its typographic choices, including
■ Body text typography (CSS)
■ Logo (if not preexisting)
■ “Type GIF” headlines, subheads, and so on

■ Navigational typography (menu bar)
Body text typography is controlled with Cascading Style Sheets (CSS), a

subject so important we devote an entire chapter to it (Chapter 10, “Style
Sheets for Designers”) and still scarcely do it justice. All we’ll do here is
remind you that 99% of the Web is text, most of it intended to be read,
and that there is neither a reason nor an excuse to create hard-to-read text
on your web pages.
13 0732 CH09 4/24/01 11:22 AM Page 240
240 HOW: Visual Tools: The ABCs of Web Type
The logo, if not preexisting, will be designed in Adobe Illustrator or Macro-

media Freehand, just as it would be in print projects. All we need to say
about that is to remember to start with web-safe colors, keep your design
simple so it can reproduce at small sizes (32 x 32 web buttons, for
instance), and pay attention to the following discussion about serif versus
sans serif faces in the limited 72ppi screen environment.
Remember the VisiBone color palette we mentioned earlier in this chapter?

Download the Illustrator version and use it to develop logos and other
graphics intended for the Web.
Before copying Illustrator artwork to Photoshop, convert to RGB via Illus-

trator’s Filter, Colors menu. The process is not perfect; web-safe colors may
shift, and you might need to select large areas in Photoshop and refill them
with web-safe colors.
The main thrust of our look at typography will not be body text (covered in
Chapter 10) or logo design (covered previously in this chapter). Instead, we
will discuss the basics of using Adobe Photoshop and ImageReady to cre-
ate typographic GIFs for the Web. We’ll also further examine how anti-
aliasing can work for or against your web designs.
THE ABCS OF WEB TYPE

As you know, Photoshop and ImageReady let you add horizontal and ver-
tical type to any image. As of Photoshop 5.5, you can specify the typeface,
leading, kerning, tracking, baseline style, size, and alignment of the type
and edit its characters. Photoshop 6 improves on its predecessor’s already
remarkable power.
Previously, such details as leading, kerning, and tracking were the exclu-
sive province of Illustrator, and most serious web designers would create
their typography in Illustrator and then cut and paste it into Photoshop.
Some still do that, and you might prefer to as well. Illustrator offers useful
keyboard shortcuts for kerning and other typographic functions. Many of
those keyboard shortcuts are missing from Photoshop, making the process
a bit less streamlined.
13 0732 CH09 4/24/01 11:22 AM Page 241
But keyboard shortcuts aside, Photoshop has advanced tremendously in its

handling of type and now offers essentially the same typographic func-
tionality that Illustrator does. As a result, many designers use Photoshop
for everything.
Photoshop 5.5 and higher also allows you to select an anti-aliasing option
for type, apply simulated styles to type, and turn off fractional character
widths to improve the appearance of small, bitmapped type displayed at
low resolution.
Anti-Aliasing
As all designers know, anti-aliasing enables you create the appearance of
smooth-edged type by partially filling in the edge pixels with intermediary
colors. For those who don’t know, we provide the following handy exercise.
Exercise 4: The Great Intermediary
Launch Photoshop and create a new blank document with a white back-
ground. Work at 72ppi. (We always work at 72ppi on the Web.)
Select the type tool. Click in the image to set an insertion point.
Enter some text in the Type Tool dialog box (Photoshop 5.5) or directly on the
image (Photoshop 6). Format the text however you like. For the sake of argu-
ment, we’ll type our names in black, 24pt. Helvetica. “Crisp” anti-aliasing is
chosen by default. (If it is not, choose it now.)
Close the Type dialog.
Go to Photoshop’s Navigator menu and blow up the image by 400%. Look at

the edges of any letter. Those soft gray pixels are anti-aliasing. Now you know.
The purpose of anti-aliasing is to fool the eye into seeing type as smoothly
rounded in spite of the low resolution of computer monitors.
Anti-aliasing is also used for images unless you’re deliberately going for a
bitmapped, pixellated look. And you’re usually not. Whether for type or
images, it can cause problems when working with GIF transparency.
13 0732 CH09 4/24/01 11:22 AM Page 242
242 HOW: Visual Tools: The ABCs of Web Type
Exercise 5: Match 'Em Up
Open Photoshop and create a new blank document with a white background.
Choose any two web-safe colors from the Photoshop Color Picker or the Vis-
iBone web palette. For the sake of argument, we’ll choose a dark purple and
a light green.
Select a circular area and fill it with the foreground color (dark purple).
Save the image as circle.psd.

Hide the Background layer so that it becomes transparent.
Save for Web.
Choose GIF (choosy mothers choose GIF) and click the Transparency checkbox.
Select the background (light green) color as your transparency color.
Optimize at 16 colors with dithering on and the web-safe slider dragged to

about 40% web-safe.
Save the image as circle.gif.
Open BBEdit or your HTML editor of choice.
Create a new basic HTML document with a background color to match the
light green (transparent) background of your GIF image.
Save the file as circle.html.
Open it in any web browser.
The circle should look good and should have a soft edge thanks to anti-
aliasing.
Return to the HTML document and change the <BODY> background color to
a new, contrasting color. Say, black (#000000).
Save the file and reopen it in the web browser.
The circle should be surrounded by an ugly light green halo.
That is improper anti-aliasing. What have we learned? Always anti-alias

against the color you expect to use in the finished web page.
How do you anti-alias a transparent type (or image) GIF when the site uses
a gradient background image or a random texture?
You can’t. So avoid using those types of backgrounds unless you never need
to set transparent GIFs in the foreground.
13 0732 CH09 4/24/01 11:22 AM Page 243
You should avoid gradient background images anyway because they will
dither horribly on 256-color monitors, don’t render properly in the GIF for-
mat, and if exported as JPEGs cannot be web-safe.
And you should avoid busy random textured backgrounds as well because
they are generally hideous, and they make text harder to read. Even beau-
tiful pages developed with subtle background tiles are not much use if no
one can read the text they contain.
The PNG format mentioned earlier offers real transparency, which means a
PNG image could be used against any type of web background without ill
effect. But the trouble with PNG is…well, we’ve covered that to death
already.
Specifying Anti-Aliasing for Type

Anti-aliasing options in Photoshop and ImageReady allow you to choose
from three levels of anti-aliasing to modify the appearance of type online.
You can choose to make type appear crisper, smoother, or heavier.
Exercise 6: Shape Up—Sizes and Faces
Create a new type layer by typing in a new, blank Photoshop document.
In the Type Tool dialog box, select an anti-aliasing option from the pop-up
menu. Choose:
■ None to apply no anti-aliasing. Useful for bitmapped fonts such as Joe

Gillespie’s Mini 7 (www.wpdfd.com/mini7.htm), Jason Kottke’s Silkscreen
(www.kottke.org/plus/type/silkscreen/), or the Fountain Type Foundry’s
Sevenet (www.fountain.nu).
■ Crisp to make type appear sharp. This is the default setting. It renders
well and uses less bandwidth than Strong or Smooth.
■ Strong to make type appear heavier. This is an impressive setting, but

because it requires additional anti-aliasing to create its effect, it fights
the LZW compression algorithm and results in larger file sizes. We are
talking about very small differences here, but these differences do
add up.
■ Smooth to make type appear, well, smoother.
13 0732 CH09 4/24/01 11:22 AM Page 244
244 HOW: Visual Tools: General Hints on Type
Experiment with different sizes and faces to get a feeling for which type of
anti-aliasing is appropriate for each face, size, and weight. This also varies
depending on the background being used, the visual interaction of other ele-
ments on the page, and so on. Most web designers choose Crisp most of the
time.
General Tips
As just mentioned, the smoother or heavier the anti-aliasing, the greater
the number of edge pixels in various shades, and the more bytes the result-
ing GIF image will require. When bandwidth is at a premium—and it is
always at a premium—err in the direction of Crisp.
Not all type needs to be anti-aliased. Smaller type might be easier to read
with no anti-aliasing at all. For instance, 10px Helvetica will be easier to
read (and will use up less bandwidth) if you choose “None” in the Anti-
Aliasing dialog box. But rather than create GIF type of that nature, a more
responsible course would be to use HTML and CSS to create small bits of
web type because such text may be easily copied, pasted, and indexed by
search engines—whereas type GIFs are simply images.
GENERAL HINTS ON TYPE

Pardon the pun. (Get it? Type? Hints? Never mind.) Every aspect of web
design involves trade-offs and potential problems for some web users.
When setting typography for the Web, here are some points to keep in
mind.
The Sans of Time

Let’s just get it over with: Sans serif fonts are far easier to read onscreen
than serif fonts. This is the exact opposite of what is true for books. But
printing is high-resolution; the computer screen is low-resolution. There
are simply not enough pixels to correctly render the tiny details required
by serif typefaces. This is especially true with smaller type, such as body
text and subheads. (It is also true for CSS text.)
13 0732 CH09 4/24/01 11:22 AM Page 245
It helps to think of your type GIFs as icons, which must be rendered pixel
by pixel in a 72ppi environment—because that is essentially what they are.
Anti-aliased fringe colors must use up an entire pixel (there are no half-
pixels). Now add subtle ascenders and descenders to this mix, attempt to
wedge such nuances into discreet pixels, and you can see why serifs work
poorly onscreen.
You also can see why typographic colors should be web-safe. Add dither-
ing to the unholy mix of anti-aliasing and serifs, and you have an illegible
mess.
This inherent preference for sans serifs on the Web might be behind the
present resurgence in Helvetica. We could be talking through our hats, but
we haven’t heard a better theory, and as we’ve shown earlier, web styles
have been entering mainstream media as fast as designers could rip each
other off.
From this discussion, it might seem that the Web is no place for fine typog-
raphy. But that is not the case. Juxt Interactive is one agency that creates
superb type treatments online, and their work repays careful study
(www.juxtinteractive.com).
Space Patrol
In most cases, web type is more readable when it is widely spaced because
such spacing makes allowances for the imprecise spreading of unruly edge
pixels. So when setting type, try loosening your tracking in the Type dialog
box. If you’ve done any TV design, it’s pretty much the same thing. If you
haven’t, just trust us.
Lest We Fail to Repeat Ourselves

Always start with web-safe colors for your type and your background to
avoid ugly dithering in low-end monitors.
13 0732 CH09 4/24/01 11:22 AM Page 246
246 HOW: Visual Tools: Navigation
Accessibility, Thy Name Is Text

The more text you create graphically, the less a search engine will under-
stand about your client’s web page and the more problems you create for
readers with disabilities or those using alternative web browsers.
As mentioned elsewhere in this book, use <ALT> and <TITLE> attributes in

your HTML <IMG> tags to explain what the search engine and the disabled
visitor cannot see. If HTML and a text GIF look equally good, choose HTML
because it increases the accessibility and usability of your page, makes it
easier for search engines to locate the relevant information, and almost
invariably uses less bandwidth than graphics.
In most cases, HTML text can be resized by the user. Type GIFs cannot. Keep
in mind that small type that looks great to you might be difficult or impos-
sible for folks with impaired vision to read.
If you were wondering why you see so much large bold sans serif typogra-
phy on the web, now you know. It’s not that web designers are copycats.
Well, we are, but it’s not just that. It’s that we’ve learned by experience that
small fonts, sans serif fonts, and tightly kerned text can all be problematic
for the people who use our sites.
As support for CSS improves, it becomes a little easier to sell clients on

CSS-style text instead of type GIFs. But resistance to this notion is wide-
spread because clients seek branding, and designers like creating it. And,
most of the time, type GIFs just work better for that purpose, regardless of
their accessibility issues.
NAVIGATION: CHARTING THE VISITOR’S

COURSE
We covered the guiding principles of navigation in Chapter 3, “Where Am
I? Navigation & Interface.” And in Chapter 7, “Riding the Project Life Cycle,”
we learned that developing a branded, intuitive navigational menu—or a
series of hierarchical navigation menus—is only the beginning and that
most web firms perform interface testing, asking volunteers to work with
the developing site. And as problems are identified, the designer is asked
to rethink and redesign.
13 0732 CH09 4/24/01 11:22 AM Page 247
Focus group testing in advertising often results in mediocre campaigns, but

focus group testing of a web interface can result in a better site—if those
who run the tests know what they are doing.
What this means in the context of Photoshop is that you will be creating
a lot of comps until you truly crack the interface problem, and then you
will be refining your comps based on feedback from user tests.
When the perfect interface has been designed in Photoshop, there is still
more to do. Often, the design team will implement a menu bar that changes
state via JavaScript as the visitor navigates the site. On the simplest level,
changing state means that the menu bar subtly indicates where the user
is within the site structure. For instance, when the visitor reaches the About
section of PlanetRX (http://www.planetrx.com/information/about.html),
the About portion of the menu bar is highlighted to remind the visitor “you
are here.” Refer back to Chapter 3’s Figures 3.2 and 3.3 to see how this “you
are here” state change is handled on the Gap site.
Changing state to reinforce the visitor’s position within the site can be
accomplished by simple HTML, via JavaScript, or with the help of publish-
ing systems that swap visual elements on-the-fly. The choice of imple-
mentation varies by the scope of the site and the size of the budget. On a
small site, the menu bar can be changed via HTML or JavaScript. On a very
large site that is constantly updated, a publishing system will probably be
used.
No matter how the technique is implemented, it is up to the designer to

create the alternate state graphics on separate layers in the Photoshop
document. (These will come in handy later in the process when the work is
sliced and produced in ImageReady.)
Typical navigation menus also “light up” or otherwise change state when
the user drags the mouse cursor over a given menu item selection. Again,
this is accomplished via JavaScript, and again, though there is no substi-
tute for home-cooked code (or working with good developers), ImageReady
can help out, as we are about to see.
13 0732 CH09 4/24/01 11:22 AM Page 248
248 HOW: Visual Tools: Slicing and Dicing
SLICING AND DICING

Photoshop is the primary tool used to design navigational menus and their
associated text (unless these menus are created in CSS, per the preceding
discussion). Photoshop and Illustrator are also used to create assorted nav-
igational elements such as arrows and buttons. The larger and more com-
mercial the site, the greater the pressure to create uniquely branded
elements.
These elements can be created in separate image documents. For instance,

you might create hundreds of arrows in Illustrator before choosing one for
your design. Similarly, you might (and probably will) go through several
rounds of logo development.
But after they are created and chosen, all of these elements are generally
layered into a single Photoshop comp, which is used to sell the work to the
client (see Figure 9.13). Of course, as we’ve just said (and as Chapter 7
explained), this “selling” is a multistage process, with continual refinement
occurring based on research, user testing, and the client’s strange whims.
Figure 9.13
Here is a Photoshop web
layout that combines
photography, logos, and
interface elements. We
used this layout to sell
a final web design to
JazzRadio.Net
13 0732 CH09 4/24/01 11:22 AM Page 249
After it’s sold, production begins, and at this point Photoshop’s ImageReady
module comes into its own. Knives were made to slice cake, and
ImageReady was made to slice web comps. The process begins by dragging
Photoshop guidelines across any area that will have to be sliced—for
instance, dragging guidelines to separate one menu bar item from the next
(see Figure 9.14).
Figure 9.14
The next phase is dragg-
ing Photoshop’s guides
to mark areas to be con-
verted to slices in the
ImageReady module.
(Photoshop 6 can create
the slices itself.) Though
slicing such comps is the
normal next step, for this
project we opened a text
editor and re-created the
layout in HTML and CSS to
minimize bandwidth and
enable the layout to
squash or stretch in true
“liquid” fashion.
With Photoshop 6, you can create and name slices right in the Photoshop
program itself. With Photoshop 5.5, having dragged guides, you “Jump to
ImageReady” via the File menu and automatically convert your guides to
slices at the touch of a button. ImageReady generates the relevant HTML,
animations (if any), and JavaScript rollover functions (if any). We don’t
mean to imply that this happens instantly, of course. There is a great deal
of typing, dragging, and layer selection involved.
Rollovers are created by selecting new layers for each rollover state and
typing the relevant URL and text (if any) in the Slices dialog box. Now you
can see why rollover states are visually designed during the comping phase.
Not only does this satisfy the client, it also enables you to focus on pro-
duction tasks without worrying about previously unconsidered design
issues.
13 0732 CH09 4/24/01 11:22 AM Page 250
250 HOW: Visual Tools: Slicing and Dicing
Performing all these production tasks is a fairly straightforward process,

and the Photoshop manual spells it out so completely that we won’t bother
doing so here. One thing Photoshop’s manual does not emphasize (but we
will) is that you can often replace selected slices with bandwidth-friendly
HTML and CSS equivalents. For example, instead of generating a large
brown GIF image, you can generate an empty table cell filled with the
appropriate background color, merely by choosing No Image from the Type
drop-down menu.
This by no means converts a browser-centric, brand-heavy site into a light,

accessible one. It does, however, help reduce overall file size, and it does
make life a bit easier for those using nontraditional browsers, given that
this will be one less pointless image to trouble them with its incompre-
hensibility.
After the process is completed, sophisticated web designers take the HTML
and JavaScript generated by ImageReady, open it in an HTML text editor
such as BBEdit, PageSpinner, or HomeSite, and edit as needed. For exam-
ple, you might substitute a simpler JavaScript function for one generated
by ImageReady.
ImageReady’s JavaScript is verily a two-edged sword. Novices and experi-

enced web designers in a hurry can rightly consider ImageReady’s auto-
mated scripting a godsend. But it is equally easy to generate massively
confusing or even completely dysfunctional scripts until you familiarize
yourself with the process. The first time we used ImageReady to automat-
ically generate image rollovers, we ended up with a folder full of bizarrely
named duplicate slices and a script that changed every image on the page
at the slightest movement of the mouse.
Then we read the manual.
Most professionals will use ImageReady to generate slices and raw HTML,
then tighten up its markup for better standards compliance and lower
bandwidth, and replace its often complex scripts with simpler ones. In large
web agencies, web technicians will perform these tasks.
13 0732 CH09 4/24/01 11:22 AM Page 251
THINKING SEMANTICALLY
Photoshop and ImageReady perform vital tasks splendidly, but what they
cannot do is generate semantic websites predicated on the separation of
style from content. Being visual tools, they necessarily create visual sites—
and of course this is what most clients want and what most designers are
comfortable with. But this is not the only way and not necessarily the best
way to create websites.
Visual sites are a comforting link to the past, to our history of print and
package design—of concrete objects made beautiful and intelligible
through precise design. Semantic sites are something else again.
Because they are rooted in images, and images are necessarily of fixed and
specific sizes, Photoshop and ImageReady generate image-laden sites laid
out in HTML tables with specific heights and widths. They do not generate
the Liquid Design we discussed in Chapter 2 because it is not in the nature
of a pixel-based program to develop abstractions of form. And certainly
they cannot separate style from content because style is their content.
So separating style from content becomes your job, if you choose to accept
it. As an interim step, what we’ve done in our shop over the past two years
is confine ImageReady’s slicing skills to key elements that must be precisely
sized—for instance, to branded navigational menu bars. But whenever pos-
sible, instead of slicing entire comps to create precise graphic web layouts,
we use our comps as guidelines to create HTML (or, even better) CSS equiv-
alents that are loose, flexible, and fairly minimalist.
This process enables us to create templates that function as “content con-

tainers.” Such sites are still branded and still function as all sites function,
but they are less tied down by fixed elements than traditional sites. This
makes them easier to revise and update (just change a style sheet) and
harder for clients to screw up when they take over the maintenance chores.
It also makes them easier for nontraditional browsers to process and posi-
tions them for the next phase of web development.
13 0732 CH09 4/24/01 11:22 AM Page 252
252 HOW: Visual Tools: Thinking Semantically
We have now broached the vital next step in the web’s history: the sepa-
ration of style from content. Meanwhile, in our discussion of web typogra-
phy, we have so far avoided the specifics of coping with actual web texts
as opposed to decorative elements. So maybe it’s time to look at a tech-
nology that handles both the separation of style from content and the need
for precise typographic control of web text.
The people of earth call it CSS, and the next chapter will explain how it
works—and what to do when it stops working.
14 0732 CH10 4/24/01 1:04 PM Page 253
chapter 10
Style Sheets for

Designers
IN THE BEGINNING WAS THE WORD: without style and unadorned on a plain gray
background.
The scientists who envisioned the Web saw it as a place for reasoned dis-
course conducted through documents whose structure was as logical as
the arguments they propounded. HTML would present content and struc-
ture, and the browser (or User Agent) would interpret it visually, according
to its own built-in rules of display. <h1>Headlines</h1> would look like
whatever the browser decided they should look like (typically, 24pt. Times).
<p>Paragraphs</p> would look like whatever the browser decided they
should look like (typically, 12pt. Times).
In early browsers such as Mosaic and Netscape 1.0, web page backgrounds
were generally gray. Why did browser developers choose this dingy color?
The answers are lost in the mists of time. In other words, we have no idea.
But we do have a theory. Namely, images seemed to want to appear against
a black background for maximum contrast and impact. Text, of course,
wanted to appear on white. We’re guessing that the makers of the first
browsers compromised by giving us a washed-out gray that would provide
rudimentary contrast for either type of foreground element. Regardless of
their reasons, the resulting web pages were not much to look at.
14 0732 CH10 4/24/01 1:04 PM Page 254
254 HOW: Style Sheets for Designers: Tag Soup and Crackers
TAG SOUP AND CRACKERS

Designers and their clients, however, were not about to sit still for such lim-
ited presentational capabilities, so browser companies (mainly Netscape at
first) began “extending” HTML willy-nilly to offer web designers more con-
trol over the visual appearance of their sites. Netscape extended the
<BODY> tag, allowing us to choose background colors as well as text and
link colors. Microsoft gave us proprietary <BODY> tag extensions that
allowed us to create margins of a limited sort.
Netscape gave us the <FONT> tag. We could control the size of our text,
regardless of its structural context. (We could, for instance, make para-
graphs really big <FONT SIZE=”7”> and headlines really small <FONT
SIZE=”1”> even if such approaches contradicted the underlying document
structure.) Microsoft gave us the <FACE> attribute for the <FONT> tag. We
could control typography in a limited, Flintstonian fashion. <FONT
FACE=”ARIAL, HELVETICA”> would make text on the page appear in Arial
if the visitor’s operating system offered that font. If not, the text would
appear as Helvetica. If neither font were available, visitors would see their
default typeface (probably Times).
While browser companies corrupted HTML in a well-meaning but wrong-

headed effort to serve designers and their clients, designers began setting
their text in Photoshop and saving the images in web-friendly GIF format.
14pt. Meta or Futura, with precise kerning and leading, looked a lot better
than <FONT FACE=”ARIAL, HELVETICA”>. Instead of using HTML to present
text, designers used it to embed visual representations of text.
What we gained in presentational spiffiness, we lost in usability. GIF images

of text could not be searched, indexed, copied, or pasted. They could not
even be seen by some people or in some browsers.
At the same time, designers began using HTML tables to control their lay-
outs, a practice most of us still follow, though it runs counter to the struc-
tural nature of HTML. The practice has another downside as well: It yokes
our presentation to our content, making it harder or even impossible for
those with disabilities or those using nontraditional browsers to access the
information on our sites.
14 0732 CH10 4/24/01 1:04 PM Page 255
Many of us went beyond using tables and text images. We harnessed invis-
ible powers to our task. As you know, in Photoshop any layer may be fully
or partially transparent. Images in the GIF format are limited to 256 (or
fewer) colors, any one of which may be designated as transparent. Using
Photoshop, web designers created small (1 pixel by 1 pixel) GIFs filled with
a single, transparent color. We then used these transparent GIF images to
control the positioning of elements on the page, as if we were designing
for a fixed medium like print.
We used these transparent GIF images again to simulate leading, inserting

“spacer GIFs” between lines of HTML text.
<IMG WIDTH=”100” HEIGHT=”100” ALT=” “ SRC=”transparent.gif”>
Notice the height and width. Netscape’s browser likes it when you indicate
the size of images used. This helps the browser leave space for the images,
even before they have finished downloading. A tangential aspect of the
whole affair is that browsers will display your images at any size you tell
them to. Thus a 1 pixel by 1 pixel transparent.gif could be 100 pixels wide
by 100 pixels tall if you marked it up that way in your HTML. These crude
feats provided rudimentary layout control, while HTML itself did not.
That was the key. HTML, practically the only tool at our disposal, provided
no typographic or layout control. So most of us deliberately deformed the
simple markup language in hopes of forcing it do our bidding. We made a
“tag soup” of the Web, using <TT> (“typewriter text”) to force the browser
to display a monospaced font (usually Courier). To create vertical space, we
deployed transparent GIFS or typed structurally meaningless carriage
returns such as:
<br><br><br>
or went so far as to create “invisible headlines” which were never intended

to be read. To create invisible headlines we used the nonstandard <FONT
COLOR> attribute to set a headline to the same color as the web page’s
background. For example, on a page whose background color was white,
we might use the following:
<H1><FONT COLOR=”#FFFFFF”>Don’t read this headline</FONT></H1>
14 0732 CH10 4/24/01 1:04 PM Page 256
256 HOW: Style Sheets for Designers: CSS to the Rescue…Sort of
By means of these stunts, the Web began to look better on the surface, but
the markup that was supposed to hold it together was becoming less and
less meaningful, more and more fragmented. Documents made less and less
structural sense and were more and more tied to the quirks of specific
browsers. “Use Netscape so you can see this page!” we screamed at our
viewers in the mid-1990s.
CSS TO THE RESCUE…SORT OF

In 1996, Microsoft, Netscape, and other members of the World Wide Web
Consortium (W3C) came up with a brilliant new standard technology—one
intended to give designers far more power over the display of web pages,
without further corrupting the structural meaning of their HTML docu-
ments. The name of that technology was Cascading Style Sheets (CSS).
CSS is the best friend a visually oriented web designer ever had, but sup-
port for this crucial standard has been a long time coming. In the follow-
ing section, we’ll gently introduce you to CSS, showing how and why it
works. Afterward, we’ll talk about what can go wrong with CSS and pres-
ent a detailed No Fault CSS Plan that enables you to harness the power of
style sheets without running afoul of buggy browsers. The good news about
all of this is that most current web browsers now offer good-to-excellent
CSS support. The bad news is that older, inferior browsers are still in use,
though they are fading away over time (see the section, “The 18-Month
Pregnancy” in Chapter 2, “Designing for the Medium,” for comments on this
topic).
As a last prefatory note, you might find yourself working at a large web
agency—one where web designers spend most of their time in Photoshop
and Illustrator, while HTML production chores are handled by a separate
group of professionals. Even at job like that, you will still need to know CSS.
Why? It’s because even when HTML chores are assigned to web technicians,
it is almost always the web designer’s job to create the style sheet.
14 0732 CH10 4/24/01 1:04 PM Page 257
That may seem puzzling. If web technicians and developers handle all other
markup and coding, why wouldn’t those professionals also be called upon
to write the site’s style sheet? The answer is simple—style sheets control
typography and layout, and that makes them the designer’s responsibility.
(You don’t really want a programmer deciding how your web typography
should look, do you?)
DESIGNING WITH STYLE: CASCADING STYLE

SHEETS (CSS)
CSS is a developing web standard whose purpose is to control the display
of web pages. Cascading Style Sheets Level 1, the initial version of CSS rec-
ommended by the W3C in 1996, is well (or fairly well) supported in current
browsers including Opera 5 or higher, Internet Explorer 5 or higher, and
Netscape 6 or higher. CSS empowers web designers to control such ele-
ments as:
■ Font families, font sizes, and leading (“line-height” in CSS-speak)
■ Margins and page divisions
■ Colors, backgrounds, whether or not backgrounds tile, whether or

not they scroll, and so on
■ Positioning of elements in relation to each other, and to the edges of

the browser window
■ Borders, HTML elements (such as <FORM> elements), and more
As this list suggests, CSS is a very powerful standard that can replace the
use of HTML tables to control layout; end the use of <FONT> tags to con-
trol web typography; and do much more than tables and <FONT> tags ever
could (see Figure 10.1).
14 0732 CH10 4/24/01 1:04 PM Page 258
258 HOW: Style Sheets for Designers: Designing with Style
Figure 10.1
The New Year’s 2001
Greeting at zeldman.com.
The background image,
text, and “core” button are
exactly positioned via CSS,
which also creates the
black outline and back-
ground colors. Notice that
the image hugs the upper
left corner of the browser
window, a feat that is
easily achieved by using
CSS to set margins and
padding at “0.” JavaScript
was used to route
Netscape 6 users, IE5
users, and Opera 5 users
to subtly different pages
(www.zeldman.com/
2001.html).
Separation of Style from Content

Beyond providing designers with a powerful tool set, CSS serves an addi-
tional purpose—that of formally separating a website’s style (or design ele-
ments) from its content (otherwise known as writing and such).
Disadvantages of Traditional Web Design

Methods
The way web designers have historically designed pages, style and content
are hopelessly intermingled. Text appears inside table cells. <FONT> tags
are wrapped around every paragraph.
14 0732 CH10 4/24/01 1:04 PM Page 259
While this old system works, and while it is used in literally millions of sites,
it has two powerful disadvantages:
1. Problems in the present: wasted bandwidth and HTML abuse.

HTML tables were never intended to be used as design tools; when
used for that purpose, they slow the rendering of web pages in the
browser and can cause problems for users of text-based browsers—
such as people with disabilities. While they do work in most
browsers, these tags and tricks slow down web pages and contribute
to bandwidth problems by forcing the user to download unnecessary
text (namely, the tags themselves). They also clutter the markup.
2. Problems for the future: retarding progress. By mingling content

with style, the present system makes it much more difficult for web
designers and programmers to create sites that can be used by non-
graphical browsers and devices, such as web phones, Personal Digi-
tal Assistants (PDAs), and audio browsers for the blind. Such devices
represent a growing and vital market. On the other hand, if content
and style are formally separated, then nongraphical browsers can
simply display text and links, while computer users with graphical
browsers still enjoy a rich visual experience created by web design-
ers. In addition to the harmful effects on web-enabled devices, the
mingling of content and style also makes it more difficult to design
and build robust interactive sites, including the e-commerce sites
you will inevitably find yourself designing.
CSS Advantages: Short Term

Under the present system, designers who wish to control the appearance
of text on the Web must type <FONT> tags for every paragraph of client
content. This adds up to hundreds of kilobytes of wasted bandwidth on
every site and hundreds of hours of tedious labor for the web designer
and/or the HTML technician.
14 0732 CH10 4/24/01 1:04 PM Page 260
260 HOW: Style Sheets for Designers: Designing with Style
After all those hours of labor, if the client requests a design change, many
more hours of labor must be put in, as the designer or web technician man-
ually searches for and replaces all those annoying <FONT> tags.
It’s a dumb way to work. With style sheets, the web designer can change
just one document—a global style sheet—and the layout and typography of
the entire site will be instantly changed. Hundreds of hours of the dullest
sort of labor can be saved in this way. If style sheets are used to control
layout as well as typography, the savings in labor (and client costs) can be
even more profound.
What can you do with the client dollars saved? You can spend them on
design, programming, writing, photography, illustration, research, testing,
marketing, and maintenance. With less of it wasted on monkey work, the
same budget now enables you to create a richer, more powerful site.
Another bonus is that after putting every ounce of our experience and tal-
ent into the design of web pages, we typically turn the sites over to our
clients, who then update the sites as needed. Websites are not carved in
stone; a site that’s not minty-fresh is a dead site. How many clients have
a background in design and extensive knowledge of web technology?
We’ve been lucky enough to find precisely one such client in nearly six
years of doing this work.
As explained in Chapter 7, “Riding the Project Life Cycle,” often during the
hand-off and maintenance phase, a junior or mid-level person with no
design experience and little web knowledge is made responsible for the
site’s maintenance and updating. Frequently, “refreshing the site” is merely
one of that employee’s daily duties. The more our pages are filled with
<FONT> tags, complex tables and framesets, the sooner that overworked
web coordinator can turn the site into an eyesore as well as a usability
nightmare. By separating design elements from content, we make it much
harder for our clients to destroy the sites we’ve worked so hard to create
for them. CSS is our friend.
14 0732 CH10 4/24/01 1:04 PM Page 261
CSS Advantages: Long Term

As indicated, CSS provides a way for web designers to create richly visual-
ized, robustly interactive sites that also might function well outside the
traditional web browser environment. As more and more people begin to
interact with the Web through new, nontraditional Internet devices—and
as more and more powerful web standards are brought to fruition in the
browser as well as at the W3C bargaining table—the need to separate con-
tent from style becomes even more important. So it’s pretty darned crucial
that web designers come to grips with this concept of style/content sepa-
ration and learn to use style sheets effectively in designing for the Web.
COMPATIBILITY PROBLEMS: AN OVERVIEW

The CSS-1 standard was created in 1996 but was not completely supported
by any web browser before the year 2000. As of this book’s publication, it
is still imperfectly supported by browsers most often used to access the
Web. This has slowed the adoption of CSS in the field given that no client
wishes to pay for a site that might not work correctly for many users.
Poor, partial, or incompatible CSS implementations in browsers also have

persuaded most web designers who do use style sheets to employ them
only in very limited ways. For example, many designers now use CSS to con-
trol the fonts on a site. But these same designers continue to use HTML
tables to control the layout of text and graphical elements on each page
(see Figure 10.2) because poor or incompatible CSS implementations in the
browser might otherwise render their layouts illegible. They can even cause
one browser to crash (more on that shortly).
14 0732 CH10 4/24/01 1:04 PM Page 262
262 HOW: Style Sheets for Designers: Compatibility Problems
Figure 10.2
The Daily Report at
zeldman.com uses CSS
to control typography but
traditional HTML tables to
lay out the page. CSS-
capable browsers are on
the market, but so are
Netscape 4 and IE3—two
old browsers whose sup-
port for CSS is problem-
atic. Because Netscape 4
users can crash from CSS
layouts and IE3 users can
barely see them, an
interim approach was
taken. When these old
browsers have faded into
disuse, the same page will
be designed entirely in
CSS (www.zeldman.com/
coming.html). We refer to this two-pronged, “safe” approach as No-Fault CSS, a tech-
nique we began recommending in 1998 in the A List Apart “Fear of Style
Sheets” series:
■ www.alistapart.com/stories/fear/
■ www.alistapart.com/stories/fear2/
The series was designed to evangelize CSS use in spite of browser compli-
ance problems by showing which CSS techniques to avoid and which could
be safely used.
Browser companies such as Netscape and Microsoft have sometimes been

slow to realize that what is good for designers and web users is also good
for browser makers themselves because fewer problems mean fewer com-
plaints and better word of mouth. Nevertheless, by fits and starts, the
14 0732 CH10 4/24/01 1:04 PM Page 263
browser companies have increasingly supported CSS in earnest. Eventually,

web designers will be able to dispense with HTML tables and other forms
of HTML abuse altogether and use CSS to design robust sites that conserve
bandwidth while offering true separation of style from content.
In turn, this separation of style from content will enable designers, pro-
grammers, and web technicians to more capably use additional web stan-
dards, such as JavaScript and the Document Object Model (DOM), to build
truly dynamic, interactive sites.
WORKING WITH STYLE SHEETS

Style sheets are composed of “rules.” Rules have two parts: a selector that
is followed by a declaration. Consider the style sheet below:
BODY {background: white; font-family: helvetica, arial, sans-serif;}

H1 {font-weight: bold; font-size: 24px; }
P, LI {font-size: 12px; line-height: 150%;}
BODY is the first selector, while the text within brackets is the declaration.
Each declaration consists of one or more properties, followed by its asso-
ciated values. For example, in the first line, background is a property, and
white is declared as its value. font-family is a property, and the fonts listed
are possible values for that property. This terminology is confusing at first,
but working with style sheets is actually very easy. Let’s look more closely
at the following example:
BODY {margin-top: 0; background: white; font-family: helvetica, arial, sans-serif;}

The first line indicates that the BODY of the HTML document will use a
white background and that typography throughout the entire page (unless
otherwise noted via an additional selector) will be in the Helvetica family.
If the user does not have Helvetica on his or her system, the type will be
displayed in Arial. If Arial is not available, a generic sans serif will be used.
Finally, margin-top: 0 tells the browser to start the web page at the top of
the browser window, rather than “helpfully” offsetting it with an unpre-
dictable or inconsistent vertical margin.
14 0732 CH10 4/24/01 1:04 PM Page 264
264 HOW: Style Sheets for Designers: Working with Style Sheets
Font families are displayed in the order with which they are written. If the
user has both Arial and Helvetica on her system, Helvetica will be displayed
because it is listed first. In this way it is possible for designers to specify
“best-case” scenarios while providing backup options. Remember: The first
font listed will be displayed if it is available. The old <FONT> tag worked
the same way.
It is crucial to provide typographic alternatives to compensate for cross-

platform differences and to end every font declaration with a generic alter-
native, such as “serif” or “sans serif.” Additional generic alternatives
include “monospace” for monospaced fonts such as Courier and “fantasy”
for ugly and cancerous fonts such as Microsoft Comic Sans.
Recognize the awesome power of style sheets. In a single line, the typeface
has been provided for an entire site. Imagine typing all those <FONT
FACE=”HELVETICA, ARIAL, SANS-SERIF”> tags instead. Yuck.
BODY {margin-top: 0; background: white; font-family: helvetica, arial, sans-serif;}
Note also that the background color has been written out as “white.” Any
color may be used, though as has been discussed before, it is always best
to use web-safe colors.
Colors need not and usually should not be specified by name because
names do not necessarily trigger web-safe colors. Instead of white, the
designer could have specified the hexadecimal code for that color: #ffffff.
It is even possible to use “shorthand” and specify only the first letter of each
hexadecimal pair (#fff). This will be clearer with a color such as #ff9900,
which can be written as #f90 in the style sheet, saving the designer three
strokes of the keyboard and saving the user an infinitesimal amount of
bandwidth.
Note that the H1 (headline), P (paragraph), and LI (list item) have had their
sizes specified in pixels:

14 0732 CH10 4/24/01 1:04 PM Page 265
It is possible (though not always useful) to specify a size of 1px or 200px

(or even larger type). Besides pixels, style sheets can use points, inches,
centimeters, .ems, percentages (“font-size: 75%;”) and even absolute font
size keywords. We will discuss the advantages and disadvantages of each
further on in the chapter. Oh, brother, will we discuss them.
Note also that it is possible to specify bold (or light, or italic, or italic bold)
and that for the first time in web design history, it is also possible to cre-
ate Quark-like leading in HTML text. Okay, you’re new to web design, so
you’re not impressed. We’ve had leading in desktop publishing tools since
Nixon wore short pants—but not on the Web, sister. On the Web, this is
some cool new stuff.
line-height: 150%
This declaration means that the text will have leading of 150%. Any num-
ber may be chosen. Line-height can be 110%, 200%, or 75% (for special
effects involving overlapping text). Assume 100% as a default, which need
not be written. (Actually, the built-in leading seems to be closer to 110%,
but again, unless you are specifying leading for a reason, leave it out to
avoid creating problems.)
Line-height, like font-size, can be specified in points, pixels, .ems, percent-

ages, centimeters, or inches.
line-height: 18px;
Because 150% of 12px equals 18px (12 + 6 = 18), a line-height of 18px

would look exactly the same as a line-height of 150% on 12px type. If the
font-size were 24px, then 150% would yield a line-height of (24 + 12) 36
pixels.
It is also possible and often desirable to indicate font-size and line-height

in the same declaration, using CSS shorthand:
P, LI {font: 12px/18px;}
The first number (12px) is the font-size; the second (18px) is the line-
height. All CSS-capable browsers understand this shorthand.
14 0732 CH10 4/24/01 1:04 PM Page 266
Leading on the Web serves exactly the same purpose as leading in print: it
adds air to the “page” and enhances readability. On the screen-based Web,
with its low typographic resolution, anything we can do to encourage read-
ability is all right by us. By contrast, reading may be discouraged when we
fail to apply leading and other CSS niceties to our text (see Figure 10.3).
Figure 10.3
The Adobe web column,
written by your humble
author, but laid out by
Adobe’s online design
team. CSS is used to
control typography, but
the small text is tough on
the eyes. CSS leading and
a larger font-size would
make the reading experi-
ence more pleasurable.
This typographic approach
works well for image
captions, a staple of the
Adobe site, but it is less
well-suited to longish
articles (www.adobe.com).
Now that we’ve taken a brief look at the rudiments of CSS, let’s see how
web designers can make this work on a site.
Types of Style Sheets

There are three main ways to use style sheets on a website:
1. By linking to an external style sheet from the HTML document
2. By embedding a style sheet within each HTML document
3. By adding styles inline in an HTML document
14 0732 CH10 4/24/01 1:04 PM Page 267
Additionally, it is possible to import one style sheet into another. Unfortu-

nately, this technique is not supported by Netscape Navigator 4, so we will
confine our discussion to the first three items. If Netscape 4 has gone to
its reward by the time you buy this book, you can read up on “CSS import”
at www.w3.org/Style/CSS/.
External style sheets

Linking to an external style sheet enables you control multiple web pages
(or even an entire site) using a single CSS document. The more pages con-
trolled by the same CSS document, the easier it becomes to make design
changes to that site. It is literally possible to change the appearance of a
5,000-page website in under a minute, simply by editing one external Style
sheet. Trust us, this is one maintenance chore you will genuinely enjoy.
Five steps to paradise: creating an external style sheet
1. Essentially, in BBEdit, PageSpinner, HomeSite, or any other HTML

editor, the designer creates a style sheet. For simplicity’s sake, here
is a basic one:
BODY {background: white; font-family: helvetica, arial, sans-serif;}
2. The designer saves this document with a filename ending in .css. For
instance, the name could be style.css, or clientname.css.
3. This CSS file is then uploaded to the server via FTP, just like an HTML
file.
4. Next, in the website’s HTML files, the designer inserts one additional
line of code within the <HEAD> tag:
<html>
<head>
<title>Welcome to Widgets.com</title>
<link rel=”stylesheet” HREF=”style.css” TYPE=”text/css”>
</head>
<body>
…and so on.
14 0732 CH10 4/24/01 1:04 PM Page 268
5. The <link> tag calls up the separate syle sheet file (style.css) and uses
its contents as instructions for displaying the page.
Note that it is possible to link to a CSS file using a relative path (“../styles/
style.css”), a rooted URL (“/path/from/server/root/style.css”), or a full URL
(http://www.widgets.com/styles/style.css). This style sheet will now control
any web page that links to it via the additional line of code within the
<HEAD> tag. An entire site can be controlled in this way.
Embedding a style sheet

Web designers who wish to affect only one page may do so by embedding
a style sheet within the <HEAD> tag of that web page.
<html>
<head>
<title>Style Sheets Rule!</title>
<style type=”text/css”>

</head>
Note the use of commenting to prevent older, non-CSS-capable browsers

from being confused by the code and to prevent search engines from point-
lessly indexing your style sheet:

What else is new in this example? The CSS is preceded by a tag that tells
the browser how to interpret it:
A more complete heading tells the browser not only that what follows is
an embedded CSS, but also tells what type of media the CSS is intended to
support:
<link REL=”StyleSheet” HREF=”style.css” TYPE=”text/css” MEDIA=”screen”>
14 0732 CH10 4/24/01 1:04 PM Page 269
The idea is that a document can link to several style sheets. For instance,
one controls screen presentation (MEDIA=”screen”), another printing, and
a third audio browsers. Not all browsers support this as of now, though it’s
a good idea to begin fully spec’ing your CSS anyway.
In a Class by Itself
All of the preceding is straightforward, but what does .sub mean in this line?
.sub {font: large bold arial, helvetica, sans-serif; margin-top: .25in;}
The selector labeled .sub is a unique class, created by the web designer for
his or her own particular design needs on this page.
That’s correct. CSS not only gives web designers the power to style traditional
HTML markup, it also enables them to create and style unique items to suit
their needs.
For instance, here, the web designer wished to create a special subhead class
with a quarter-inch margin at the top. She decided to call it sub because the
name was easy to remember and indicated the purpose (subhead) for which
the class was created. The designer could have called this class unclecharlie
if she wished.
To make use of this special class, the web designer will refer to it in the HTML
document in this way:
<div class=”sub”>Here is my subhead, with a quarter-inch margin at the top.</div>
In the web page, the line, Here is my subhead, with a quarter-inch margin at
the top would be large, bold, Arial or Helvetica (or generic sans serif) with
(surprise!) a quarter-inch margin at the top.
Style sheets rock.
Adding styles inline

The inline method is used when the web designer wishes to change the
appearance of a single tag or group of tags on one page, and not for chang-
ing the entire page or site. Adding styles inline does not offer web design-
ers the true power of CSS because it forces them to restyle text one item
at a time. Still, it can be very useful at times.
14 0732 CH10 4/24/01 1:04 PM Page 270
For instance, an entire page or site may be set in medium-size Verdana

(Helvetica, Arial, sans serif). But one line of text needs to stand out from
the rest. Perhaps this line of text represents a letter from a customer—or a
message from the U.S. Internal Revenue Service. The web designer decides
that this particular line of text should be set in 12px Monaco.
She could create an entire class just for that line of text and include that
class in the site’s global style sheet, but why create an entire class for one
line of text on a single web page? Inline styling does the job better:
<p style=”font: 12px monaco, monospace;”>
Greetings from the I.R.S.</p>
Inline styling seems like an oddity, but it is actually a wonderful supple-

mental tool—much like a tube of touch-up paint that is used to correct a
small detail.
Inline styling is also quite effective for improving the appearance of

<FORM> elements:
<div align=”center”>
<form>
<input
type=”button” style=”font-size: 12px; font-family: geneva, arial; background-color:
#ff6600; color: #ffffff;”
onClick=”window.location=’com0400d.html’;”
</form>
</div>
Form elements also may be styled via DIV classes in a global style sheet. If
every <FORM> button on your site is supposed to be orange (#ff6600) and
use 12px Geneva or Arial type, by all means create an orangebutton
class for the site, declare it on the global style sheet, and then refer to it in
individual HTML pages, like so:
<div align=”center”>
<form>
<input
type=”button”
class=”orangebutton”
14 0732 CH10 4/24/01 1:04 PM Page 271
onClick=”window.location=’com0400d.html’;”
</form>
</div>
TROUBLE IN PARADISE: CSS COMPATIBILITY

ISSUES
The first web browser to attempt to support CSS was Microsoft Internet
Explorer 3.0 (1997). It supported about 30% of the standard. A year later,
Netscape 4 came out with support that was marginally better than that of
IE3. During three years of hell, while Netscape sought to rebuild its browser
from the ground up, Navigator 4 sat rotting on the market—its once-proud
CSS support looking more and more shoddy. IE4 got more of it right and
was soon replaced by IE5, which got still more of it right. No browser got
it absolutely right, and baffled web users were often reluctant to upgrade
to incremental (4.52 anybody?) versions of these browsers.
Thus support for CSS lagged, and problems abounded. In 1998, The Web
Standards Project (www.webstandards.org) was formed to advocate and
shore up support for CSS and other web standards, and A List Apart
(www.alistapart.com) began running the “Fear of Style Sheets” series.
Fear of Style Sheets: CSS and Layout

One of the great strengths of CSS is its ability to position elements on a
web page. Elements may be positioned by exact pixel coordinates (400 pix-
els from the top, 32 pixels from the left, for example). They may be posi-
tioned relative to each other. They may be positioned via percentages,
permitting web designers to easily set up liquid layouts, as previously
discussed.
Unfortunately, CSS positioning is not supported in IE3 and is poorly

supported in Netscape Navigator 4. In fact, it can cause the browser to
crash, as detailed in A List Apart’s article, “The Day The Browser Died”
(www.alistapart.com/stories/died/).
14 0732 CH10 4/24/01 1:04 PM Page 272
272 HOW: Style Sheets for Designers: Trouble in Paradise
So until IE3 and Netscape 4 leave the market, many of us will probably keep
using HTML tables to create our web layouts. Web pages laid out with
HTML tables will work in Netscape 1, 2, 3, 4, and 6; in IE 2, 3, 4, 5 and
beyond; and in all other graphical browsers, including Opera and iCab.
If you’ve been following along, you’ll realize this means that web design-
ers still cannot safely separate style from content on commercial projects
and will continue to face difficulties in creating sites that work well out-
side the traditional desktop computer-based browser. But in the trenches,
where work gets done, it also means that we can create sites that work for
our clients and our clients’ audiences.
By late 2001 or soon after, we should be free to truly harness the power of
CSS. Meanwhile, on personal, noncommercial projects, we can explore the
full potential of CSS and other web standards without fear of hurting our
clients’ customers (see Figure 10.4). It is hoped that these noncommercial
usages of CSS and other web standards help widen interest in emerging
technologies and encourage quicker adoption of newer, more standards-
compliant browsers.
Figure 10.4
Web Trumps, a card game
featuring well-known web
personalities, uses CSS to
control the positioning
and layering of every
image on the page and
JavaScript to reveal new
layers during game play.
Web Trumps is a mini-
masterpiece of graphic
design and web program-
ming—but one requiring
the use of a modern
browser. Experiments like
this help hasten the day
web designers can apply
the same level of sophisti-
cation to commercial
designs without worrying
about browser incompati-
bilities (http://
pixelflo.com/008/).
14 0732 CH10 4/24/01 1:04 PM Page 273
Fear of Style Sheets: Leading and Image

Overlap
As mentioned earlier in this chapter, CSS leading provides a standard
means of improving the legibility and aesthetics of textual presentations
on the Web—something every site designer should care about. And it does
this while avoiding all the problems associated with transparent GIF hacks
and other nonstandard visual workarounds.
The CSS line-height property solves all the old problems, but it can lead to
new ones, particularly in older browsers whose support for CSS is largely
theoretical. For instance, in the following example, if CSS line-height is
specified for the <p> tag, the image will float on top of the text in both
IE3 and early versions of Navigator 4:
<p><img src=”dog.gif” width=”100” height=”100” alt=”My dog, Pookie.” title=”Pookie is a

friendly mutt.” align=”left”>My dog Pookie liked this text so much, he decided to sit on it
in IE3 and Navigator 4.</p>
This problem is more prevalent in Mac OS than in Windows, largely because

big browser companies spend more time and resources developing
browsers for Windows than for other operating systems. (All the more
miraculous then, that in the year 2000 the best browser on the market was
IE5/Mac. Arguably it is still the best.) The solution to image overlap is to
keep images outside of paragraphs and <div> tags. Unfortunately, doing so
brings up yet another problem in Netscape Navigator 4:
<img src=”dog.gif” width=”100” height=”100” alt=”My dog, Pookie.” Title=”Pookie is a

friendly mutt.” align=”left”>
<p>My dog Pookie liked this text so much, he decided to shove it to the right in its own
little column in Navigator 4.</p>
With a left-aligned image placed outside the <p></p>, some versions of

Netscape Navigator 4 will stick all the paragraph text in an imaginary col-
umn to the right of the image, as if the web designer had desired to cre-
ate such a column. New paragraphs, in turn, will appear also in that
unwanted column. Many a simple layout has been ruined in this way.
There is no solution to this problem. No matter which approach is used,

some Netscape 4 users are going to get an ugly layout. Because an
unwanted and mindlessly stupid column is preferable to text that is hid-
14 0732 CH10 4/24/01 1:04 PM Page 274
den behind an image, it is best to hope and pray that most Netscape 4.x
users are equipped with a more recent version of the browser. Alternately,
the designer can create pages that use no images whatsoever—scarcely an
attractive option. Finally, the designer can wrap images inside table cells,
given that doing so seems to solve most of these problems—at the terrible
cost of adding needless, bandwidth-wasting and time-consuming code to
every single web page.
The good news of course is that Netscape 6 avoids all these problems, and
Netscape 4, like other old browsers, will gradually wither away. The bad
news is it hasn’t withered away yet. So proceed with caution.
Fear of Style Sheets: CSS and Typography

Guerrilla warfare pays little heed to niceties and neither can designers in
the trenches. Too much of CSS still does not work in millions of web users’
browsers. To prepare you for battle, we will now pay little heed to the way
things should work. Instead, we will show you what does work in any CSS-
capable browser, no matter how old, inadequate, or semi-standards-
compatible it is. In other words, the following is an interim strategy for use
until nearly all web users are packing a CSS-compliant browser. If you wish
to control your web typography with CSS (and why wouldn’t you wish to
do that?), there are only two things that always work:
1. Use pixels (not points, not .ems, not percentages, not keywords) to
specify your font sizes.
2. Or use nothing. Do not specify font sizes at all, and let the browser’s
stylistic defaults and the visitor’s preferences take care of the rela-
tive size relationships. This approach is detailed in the “Dao of Web
Design” article at A List Apart (www.alistapart.com/stories/dao/)
Promise and performance

By now you understand that CSS is an important standard. It allows web
designers to specify the font family, size, margins, and leading of type on
the web; permits web designers to create advanced web layouts without
abusing HTML; and enables web designers, web practitioners, and pro-
grammers to separate design elements (presentation) from content.
14 0732 CH10 4/24/01 1:04 PM Page 275
This ability to separate presentation from content theoretically empowers

us to create attractive sites without excluding visitors who cannot use
graphical browsers—a highly desirable goal. It also paves the way for the
expansion of the Web beyond the desktop computer and onto a variety of
hand-held and other Internet-enabled devices.
Yet many times our best CSS efforts fail in one browser or another.
Even though the CSS Level 1 standard was finalized in 1996, the first
browser to meaningfully support it did not appear until the year 2000
(Internet Explorer 5, Macintosh Edition). Fortunately, Netscape 6 (multiple
platforms), Opera 5 (multiple platforms), and Konqueror (Linux/UNIX) soon
followed, with commendable CSS support of their own. But each of these
fine browsers enjoys only a relatively small market share as of this writing.
At present, the market is dominated by IE for Windows—a browser that

comes teasingly close but misses the mark in a few critical areas. The Win-
dows version of Microsoft’s browser did not fully support CSS-1 before the
release of IE6—if then. And Netscape Navigator 4, still used by tens of mil-
lions, does such a poor job of handling style sheets that it has been known
to crash upon encountering them, as detailed in A List Apart’s “The Day the
Browser Died.”
Faced with these inconsistencies, many web designers have avoided using
CSS altogether. A few brave souls have leaped ahead to fully exploit the
power of CSS in spite of the dangers posed to old browsers. Other web
designers and developers have followed the “No-Fault CSS” plan outlined
in A List Apart’s “Fear of Style Sheets” series, whether they picked it up at
ALA or figured it out on their own.
Still others—tricky devils—have created platform and browser detection

scripts to serve a variety of “appropriate” style sheets to specific user
agents—for instance, serving one style sheet to an IE4/Mac user and
another to a Navigator 4 user on Windows NT. This approach was always
unpleasantly complicated, but at least it used to work. As we’ll show you
in a moment, it no longer works.
What works? Pixels or no sizing at all works. How can we make this auda-
cious claim? We’ll let an expert make it for us. Take a sad look at Web
Review’s Master List and see the inconsistencies for yourself:
14 0732 CH10 4/24/01 1:04 PM Page 276
The Master List

www.webreview.com/style/css1/charts/mastergrid.shtml
“The mother of all CSS (Cascading Style Sheet) charts,” which lists every
aspect of the CSS spec and identifies how well it is supported by various Mac
and Windows web browsers.
If the Master List did not convince you (or if you could not quite grasp its
meaning), we’ll look at the alternatives one by one:
Font Size Challenges

Among many other capabilities, CSS allows web designers to specify the
size of typography on web pages. As shown below, font sizes can be indi-
cated using any of the following: points, pixels, absolute keywords, relative
keywords, length units, and percentage units.
H1 {font-size: 14pt}
H1 {font-size: 14px}
H1 {font-size: x-large}
H1 {font-size: larger}
H1 {font-size: 1.5 em}
H1 {font-size: 125%}
Too bad most of the stuff doesn’t work everywhere…yet.
Points of contention
Points are the units of measure with which designers are most familiar—
from their years of creating print layouts in Quark XPress or similar pro-
grams. Unfortunately, points are meaningless on the Web. Points function
as units of print, not as units of screen space. (Pixels are the only mean-
ingful unit of screen space.) Due to platform and resolution differences,
14pt. can mean many things. What it does not mean is a specific unit of
screen size.
Points are included in the CSS spec so that designers can set up a second
style sheet for printouts, as mentioned earlier in this chapter—one CSS doc-
ument to control the way the display looks on the screen and a second for
printing.
14 0732 CH10 4/24/01 1:04 PM Page 277
In your print-oriented style sheet (if the browser supports this), it makes
perfect sense to use points because printers understand points and can be
thrown for a loop by pixels. In some older browsers, 12px type gets printed
as 12 pixels, and those pixels are computed against the printer’s resolution.
Got a 1200ppi printer? Your 12px type could be .01 inches tall. To avoid
that kind of lunacy, points should be used in style sheets devoted to the
printer—and nowhere else. (Better browsers recalculate style sheets
according to the needs of the printer, but your visitors may not be using
these browsers.)
In the world of print, there are approximately 72 points per inch. To match
this, Mac OS offers a default resolution of 72 pixels per inch, mapping pix-
els to points (give or take a fraction). Of course, as soon as the Mac user
changes her screen resolution, all bets are off. In Windows and other PC
operating systems, there are 96 pixels per inch—until the PC user changes
her screen resolution, and then all bets are off.
What this means is that point sizes are incompatible between Mac OS and
Windows right from the get-go. For instance, when a Windows client sends
a Microsoft Word document to a Mac-based graphic designer, the type is
often too small for the designer to read. The same problem traditionally
plagued web pages.
Leaving aside the fact that most users change their screen resolution (and
therefore all bets are off), savvy developers have used JavaScript to serve
appropriate point-size-based style sheets to Mac users versus PC users. It’s
more complicated than using pixels, but at least it used to work.
Point of no return: browsers of the year 2000

In IE5/Mac and Netscape 6, this has changed. (See ALA’s “Why IE5/Mac
Matters” for a complete discussion of this issue.)
IE5/Mac sets as its default typographic preference 16px type at 96ppi. In

other words, it brings the default Windows typographic resolution to the
desktop of Mac users. Netscape 6 does exactly the same thing.
14 0732 CH10 4/24/01 1:04 PM Page 278
This is not evil hegemony; it is simply common sense in that the more
closely browsers adhere to commonalities, the less likely web users are to
get hurt. Windows’ default resolution is no better or worse than Mac res-
olution. But it is the most commonly used resolution, so more sites are
designed to accommodate it. Treating it as a de facto “standard” prevents
Mac users from being hurt by the poor authoring practices of some web
developers.
“Aha!” cries the Scripting Brigade. “So all we have to do is add a few more
lines of code to our browser detection scripts, and we can serve Windows-
size type to Mac users if they are surfing with IE5 or Netscape 6 and Mac-
size type if they are using older browsers?”
Not so fast, buckos. IE5/Mac starts at this default resolution but enables
users to change it. They can change it back to standard Mac resolution (and
how will you know if they’ve done that?). Adept users can change it to a
size based specifically on their screen resolution, and Netscape 6 users can
change their font size to any arbitrary value that strikes their fancy. You
have no way of knowing if they’ve done this or not. Therefore, using
JavaScript to detect the user’s browser and platform tells you exactly noth-
ing about their default font size and its relationship to standard point sizes.
There is only one thing of which we can be certain: If you use points to
control sizes, you are kidding yourself.
What works? Pixels.
Pixels for fun and profit

Though screen resolutions vary, a pixel is always equal to a pixel. Pixels are
the only reliable means of sizing typography if the web designer absolutely
must control the size of type on the web page. Unfortunately, this practice
might cause problems for some readers. For instance, if the designer has
specified 10-pixel type:
1. The visually impaired might have difficulty reading the type. This is
not a problem in IE5/Mac, which allows users to resize type at their
discretion by using the included “text zoom” function. Netscape 6
offers similar functionality, and Opera 5 zooms the entire page at the
touch of a button. So in those browsers, you can use pixels without
causing accessibility problems for anyone. (But these, as we’ve
already explained, are not the most popular browsers.)
14 0732 CH10 4/24/01 1:04 PM Page 279
2. Older browsers do not allow visitors to resize most CSS type—

particularly type set in pixels, and IE5.5/Windows still does not offer
text zooming at all. Thus, there will always be a potential accessibil-
ity hazard involved when you specify text in pixels—at least until IE
for Windows offers text zooming or an equivalent solution. As
explained in Chapter 2, we might have to wait 18 months or more
for Netscape users to upgrade to the 6.0 browser and for Microsoft
to implement text zoom in its Explorer browser for Windows.
3. If your style sheet calls up a scalable True Type font and if the user’s
operating system includes that font (and supports True Type), your
pixel-based style sheet will work just fine. But if the user’s system
does not include a scalable version of that font or a suitable substi-
tute or does not natively include True Type fonts (Linux for example),
type set in pixels can display jaggedly and may be illegible.
Accessibility problems are deadly serious. This is not idle, theoretical

chitchat. When people can’t read (or even access) your site, it hurts them,
it hurts you, and it hurts your clients.
Accessibility problems aside, some designers quibble that pixels are bad
because they vary according to screen resolution. A 400 x 400 pixel square
fills most of the screen at 640 x 480, and very little of it at 1600 x 1200.
To which we reply, so what? A 100 pixel–tall CSS headline will be the same
height as a 100 pixel–tall GIF image. A 200 pixel–wide CSS div will be
equivalent to a 200 pixel–wide JPEG image.
If you intend to create print-like layouts on the web—or even liquid layouts
that depend on the relative sizing of elements—you have exactly one
choice: pixels. If you can get away with a looser type of design (as you can,
for instance, in a personal diary or an academic paper), so much the bet-
ter. Most of us have to size the elements in our layouts, and most of us
designers like it that way.
Besides, our other options simply do not work. For instance:
14 0732 CH10 4/24/01 1:04 PM Page 280
Absolute size keywords

There are seven absolute size keywords in CSS Level 1:
xx-small medium large

x-small x-large
small xx-large
If implemented correctly and uniformly, these seven keywords would allow

designers to specify approximate sizes without running into the accessi-
bility problems associated with pixels. For that reason, the W3C recom-
mends their use. The W3C is wise, and the recommendation is
sound—except that it fails in too many browsers.
One size fits nobody

Unfortunately, absolute size keywords are unusable in many browsers.
Netscape 4 largely ignores them. Netscape 4.5 and higher and IE3 render
them at illegible sizes. (For instance, Netscape 4.5 and IE3 render xx-small
at 6 pixels, which is 3 pixels below the threshold of legibility.) In Netscape’s
case, the engineers were following an early recommendation of the W3C,
which was that each size should be 1.5 times larger than the size below it.
If small was 10 pixels, medium (one size larger) would be 15 pixels.
The W3C later changed its recommendation, but not before Netscape had
followed it. We can’t fault Netscape for trying to support standards that
changed, but we can point out the absurdity of using absolute size key-
words if even one of your visitors is using Netscape 4 or IE3. And millions
of folks use those browsers.
Small means medium, war means peace

Does IE5.x/Windows get it right? Not in our estimation. In IE5.x/Windows,
there is a logical disconnection between the keyword and the way it is ren-
dered. “Small” is displayed as medium; “medium” is larger than normal; and
so on. (IE/Windows gets keywords right.)
The engineers who developed IE for Windows are not hacks and are not evil.
They were trying to do the right thing. Remember the seven <FONT SIZE>
settings supported by Netscape? Sure you do—<FONT SIZE=1>, <FONT
SIZE=4>, and so on. Rather deftly and cleverly, the IE developers mapped
14 0732 CH10 4/24/01 1:04 PM Page 281
the seven CSS keywords directly to the seven Netscape font sizes. In many
ways, it was a logical and even brilliant thing to do. (The IE/Windows devel-
opers were also the first group to attempt to support absolute font size key-
words. We should credit them for that before carping about the results.)
The problem, of course, is that, logically, the sizes do not map to the key-
words. In old-style browsers, <FONT SIZE=3> is the default or normal size
that the user has specified in her preferences. In Netscape’s extended HTML
markup, <FONT SIZE=3> is assumed unless you specify another size. Logi-
cally, a default size should map to the “medium” CSS keyword. Unfortu-
nately, in the IE/Windows scheme, <FONT SIZE=3> maps to “small” instead
of “medium” because small is the third size up from the bottom of the list.
Who goofed—the W3C or the IE/Windows team? It doesn’t really matter.

What matters is that the keywords don’t map to expected sizes, and an
incompatibility exists not only between different manufacturers’ browsers
but between the Mac and Windows versions of the same browser.
If you think of the seven sizes the way the IE/Windows team did, your sizes
will be off on Mac users’ desktops. (You also will go nuts. It’s like trying to
drive a car where Park means Neutral.) If you think of them the way the
keywords actually read (small, medium, large) your display will be off in
Windows. You can trick the Mac browser into emulating Windows behav-
ior by specifying a <DOCTYPE> of HTML 4 Transitional and leaving off the
W3C URL. (For details, see http://www.alistapart.com/stories/ie5mac. But
this is forcing the browser to emulate nonstandard behavior, and that's not
good. Besides, it won't work in Netscape 4, Opera, or Konqueror.
So what can you do? Sadly, until your entire audeince uses browsers that
render absolute keywords, all you can do is ignore the W3C recommenda-
tions and use pixels in your style sheet. Or do not use sizes at all.
Relative keywords
Relative keywords are limited to two: smaller and larger. These in turn refer
to the size of the parent element. For example, consider the following
example, in which the <BODY> is 12px, and <BOLD> is “larger.”
14 0732 CH10 4/24/01 1:04 PM Page 282
<HTML>
<STYLE TYPE=”text/css”>

</STYLE>
Bold type would theoretically be 14px tall in this example because 14px is
one “notch” up from 12px. Like absolute size keywords, relative keywords
are ignored or bungled in some popular browsers (Explorer 3 ignores them,
as does Navigator 4 for the Mac). And even if they worked correctly, they
would be insufficient for the needs of most web designers and their clients.
Normal, larger, and smaller is not exactly a robust vocabulary for the needs
of professional designers.
So what can you do? You can use pixels in your style sheet; that’s what you
can do.
Length units
Length units sound smutty (those W3C folks should get out more…or
maybe it’s just us) and include the following:
■ em—Is a unit of distance equal to the point size of a font. In 14pt.

type, an em is 14pts. wide—named for the size of the capital “M.” But
you knew that.)
■ ex—Refers to the height of lowercase letters.
When used with the font-size property, em and ex units refer to the font
size of the parent element.
<HTML>

</STYLE>
14 0732 CH10 4/24/01 1:04 PM Page 283
In this example, <STRONG> would be 24px tall, or 2em (two times the font
size of the parent element, which is the <BODY> tag). In theory, a web
designer could create a layout using em or ex units, where all elements
were sized relative to each other. This would avoid the accessibility prob-
lems associated with pixels.
Unfortunately, the browsers make this impossible for the time being.
Netscape 4 ignores em and ex units. IE3 treats em units as pixels. Thus, 2em
is mistranslated as 2 pixels tall. It takes a village to raise a child, and it takes
at least 9 pixels to render a font. Length units are therefore not recom-
mended. What is recommended? Pixels or nothing.
Percentage units
Percentage units, like length units and relative keywords, refer to the size
of the parent element.
<HTML>

</STYLE>
In this example, <STRONG> would be 24px tall, or 200% of the font size
of the parent element, which is the <BODY> tag. In theory (notice how we
keep saying “in theory"?), a web designer could create a layout using per-
centages and avoid the accessibility problems associated with pixels.
Nothing is sadder than the murder of a beautiful theory by a gang of ugly

facts. Netscape 4 for Mac renders percentage units when they are used for
line-height (leading) but ignores them entirely when they are used to spec-
ify type sizes. And some versions of Netscape 4 for Windows treat per-
centages as pixels. (Thus, 200% is dementedly translated as 200 pixels.
Mmm, nice layout.)
14 0732 CH10 4/24/01 1:04 PM Page 284
Lest we forget, good old IE3 drops the ball by sizing percentages relative
to the user’s default font size rather than to the parent element. In Eng-
lish: If the web user has set her browser’s default to 10px, IE3 will display
<STRONG> at 20px and not the 24px you intended. If her browser defaults
to 16px, <STRONG> will be 32px. Too bad for you. Too bad for your visitor.
In spite of their accessibility benefits, percentages still fail in too many

browsers. What works? Pixels—pixels or nothing. In case we’ve failed to
communicate, we will summarize our findings as follows:
Looking Forward
Web designers will continue to be limited to using pixels in their style
sheets—despite the accessibility hazards associated with that practice—
until fully standards-compliant browsers exist and are widely used. The
approach might have its drawbacks, but failure to work correctly is not one
of them. As web designers, our job is to control the visitor’s visual experi-
ence to communicate. For the time being, the approach outlined here will
allow us to do exactly that. And soon enough, Lord (and browser compa-
nies) willing, the full power of CSS will be ours.
Can you take CSS further today? Quite possibly. It depends on the makeup
of your audience and your salesmanship with clients. As explained in Chap-
ter 13, A List Apart converted to an all-CSS layout in February 2001, and
many sites have since followed suit. For details and encouragement, see
http://www.alistapart.com/stories/99.
15 0732 CH11 4/24/01 11:23 AM Page 285
chapter 11
The Joy of JavaScript
WE’VE SAID ALL ALONG that the Web is not print. If you’ve harbored lingering
doubts on that score, this chapter should clear them up pronto. If this
chapter does not clear them up, try club soda and a semiabrasive cloth.
A primary reason the Web is not print is that websites don’t just sit there;
they do things—responding to clicks of the mouse, hovering motions, and
other user activities. JavaScript is behind much of that interactivity. In
JavaScript parlance, user actions such as mouse clicks are called events,
and you handle them via event handlers (“onClick,” for example). Similarly,
in JavaScript, the components of a web page, such as GIF images and form
buttons, are considered objects. Web pages are known as documents, and
the whole shebang is held together by a Document Object Model (DOM).
See, it’s not that hard, and you are learning already.
In this chapter you’ll find out what JavaScript is, where it came from, how
it brings websites to life, where to learn all about it, and how to begin using
it even before you learn all about it.
You’ll also learn how to communicate desired JavaScript functions to the

developers on your team, who will handle the heavy scripting when it is
needed. If you freelance or work in a graphic design shop instead of a web
agency, you’ll learn how to communicate with freelance programmers. But
before you can begin doing any of that, a few basics are in order.
15 0732 CH11 4/24/01 11:23 AM Page 286
286 HOW: The Joy of JavaScript: What Is This Thing Called JavaScript?
WHAT IS THIS THING CALLED JAVASCRIPT?

JavaScript is a programming language designed specifically to work in web
browsers. Its purpose is to bring interactivity to sites. Though JavaScript is
powerful and complex, it is relatively easy to learn—at least it is easier to
learn than many other programming languages.
Even before JavaScript, sites could be somewhat interactive. After all, click-
ing a hyperlink loads a new page. The nonlinear nature of hypertext allows
the reader to decide where to go next—to structure her own voyage
through the site (and, indeed, through the Web).
That is an interactive process and a rather profound one. But it is not a ter-
ribly sophisticated form of interactivity. JavaScript puts the top hat and
tails on web interactivity.
The Web Before JavaScript

Before JavaScript, programming languages such as Perl were used to facil-
itate interactive processes, for example typing text into a form and click-
ing a button, thereby sending requested information to the site’s owners.
Perl is still often used for this purpose on a great many sites.
But Perl is a server-side scripting language. That is, when a visitor clicks the
Send button, the web server itself must process the script. If the server goes
down or malfunctions, nothing will happen. Likewise, a web page not con-
nected to a web server—say, a web page on your hard drive—would not be
able to process such a script except in special circumstances (permanent
Net connection, full URL specified in the <FORM ACTION>, burning of
incense, wearing of magic ring).
While the script is being processed, the web server is momentarily tied up—
just as your Mac or PC gets tied up when you apply motion blur to an image
in Photoshop. Imagine constantly applying motion blur while receiving and
sending email, and you begin to see what web servers were up against. (This
is, of course a crude picture, and if you like it, we have other crude pictures
available at a nominal price.)
15 0732 CH11 4/24/01 11:23 AM Page 287
As the Web’s population grew, more and more users were clicking more and
more Send buttons. More and more web servers were thus spending more
and more time processing scripts in between serving web pages. Mean-
while, the user’s browser sat there, doing nothing. A tool was needed to
transfer the process from the server to the client (the user’s computer).
Enter JavaScript.
JavaScript, Yesterday and Today

In 1995, the company formerly known as Netscape Communications Cor-
poration introduced a client-side programming language designed to
transfer the burden of interactive processing from the web server to the
end-user’s browser. Unlike other programming languages (such as Java or
C), this new language was built into the browser. It even understood HTML.
Netscape called its new client-side scripting language LiveScript, but soon
changed the name to JavaScript to capitalize on the growing excitement
about Sun’s Java language. To this day, as a result of their similar names,
many beginning web designers (and users) confuse Java with JavaScript.
The relationship between the two is mainly one of marketing.
Netscape quickly promised to release JavaScript as a web standard so that

other browsers could use it too. But for competitive reasons, the company
initially held back on its pledge. Old browsers like Mosaic could not use
JavaScript at all. Neither could Microsoft’s newly unveiled Internet
Explorer.
To offset Netscape’s advantage, Microsoft’s browser engineers developed a

JavaScript-like language called JScript. Web design quickly became a cir-
cle of Hell, as serious developers were forced to work with these similar but
incompatible technologies. Freelancers and small firms, lacking deep
resources, often chose to support only one technology. That one was nearly
always JavaScript, especially in the beginning when Internet Explorer
enjoyed only a tiny share of the market. Thus JavaScript functioned as a
sort of standard before it really was one.
15 0732 CH11 4/24/01 11:23 AM Page 288
288 HOW: The Joy of JavaScript: JavaScript, Unhh! What Is It Good For?
Around the Millennium, JavaScript finally became an official web standard,

available to all. As reported in Chapter 2, “Designing for the Medium,” the
European Computer Manufacturers Association (ECMA) supervised the
standardization of JavaScript, renaming it in the process. The universal web
scripting language is now officially known as ECMAScript or ECMA-262,
though no one we know calls it anything but JavaScript. (Our Scandana-
vian friends pronounce it “Ya-va-script,” which we find incredibly endear-
ing. That has nothing to do with any of this, but it does lend this chapter
a certain international flair.)
The point is that JavaScript/ECMAScript is a standard that works in all cur-

rent browsers, though there are still some subtle incompatibilities being
worked out between the latest versions of Netscape, Internet Explorer,
Opera, and other browsers. Meanwhile, the DOM has been standardized by
W3C. Prior to that, JavaScript had its own ever-changing DOM in Netscape,
and Microsoft had its own DOM as well.
As all browsers finalize complete support for these standards, web design-
ers and developers are being empowered to create sophisticated interac-
tivity that works for everyone. Conversely, as browsers lag in their DOM and
ECMAScript support, web designers and developers are stuck programming
alternate versions of every site function, often going so far as to develop
alternate versions of the sites themselves.
Now that we’ve concluded our mini-history lesson and you’ve finished your
nap, let’s move on to the good stuff.
JAVASCRIPT, UNHH!
WHAT IS IT GOOD FOR?
Absolutely lots of things. Through this web-friendly programming lan-
guage, designers and developers can:
■ Replace cryptic status bar URLs (“http://www.doglovers.com/

poodles.html”) with text messages (“Learn about poodles”), a some-
what controversial practice, for reasons we’ll discuss in a moment.
■ Create the ever-popular image rollover effect discussed in Chapter

9, “Visual Tools.”
15 0732 CH11 4/24/01 11:23 AM Page 289
■ Compensate for browser incompatibilities.
■ Open new, precisely sized “pop-up” windows, with or without

various bits of browser chrome.
■ Test for the presence or absence of Flash, QuickTime, or other plug-

ins (more about plug-ins in Chapter 12, “Beyond Text/Pictures”).
■ Rotate content and images depending on the time of day, the num-
ber of times the user has viewed a certain page, or simply at random.
■ Enable the client to inject 50 links on a page without cluttering that

page at all.
■ Provide alternative means of navigating the site.
■ “Remember” that a user has visited the site before, sparing her the
pain of reentering personal data or passwords. This is accomplished
by means of cookies (Netscape terminology for little bits of text that
reside on the end-user’s hard drive and can be recognized by
JavaScript on subsequent visits to the site).
■ Cause images or text boxes to scroll horizontally or vertically,

another controversial and often annoying practice.
■ Verify the credibility of email addresses entered on “customer feed-

back” forms. JavaScript won’t tell you if a person is using her email
address or someone else’s, but it will tell you if the address is well
formed or not. (If malformed, it is probably a nonworking address.)
■ Control complex frames—of less importance than it used to be, as

frames are gradually being phased out. Similarly, JavaScript can pro-
tect sites from a third party’s poorly crafted frames.
■ Create nested navigational menus that reveal secondary and tertiary

levels in response to cursor movements—a wonderful idea because it
enables visitors to navigate directly to the information they seek, but
problematic because not all browsers fully support a standard means
of doing this.
■ …and much more.
Add the W3C DOM to what JavaScript does already and you can change
that phrase to “much, much, much more.”
15 0732 CH11 4/24/01 11:23 AM Page 290
290 HOW: The Joy of JavaScript: Sounds Great, but I’m an Artist
We know what you’re saying. “Sounds great, but I’m, like, an artist. Do I
really have to learn this stuff?”
SOUNDS GREAT, BUT I’M AN ARTIST. DO I

REALLY HAVE TO LEARN THIS STUFF?
The politically correct answer is, yes you do, because adding interactivity
to your clients’ sites is part of what makes you a web designer. The gentle
answer is, learning JavaScript is an iterative process: You can begin by cut-
ting and pasting and gradually come to understand what you’re working
with. The Richard Nixon Memorial answer is, not at first, and maybe never.
Not at first and maybe never is an answer because many working web
designers get by for years doing nothing more than cutting and pasting
other people’s scripts. By the way, we’re not talking about stealing code.
Many developers freely offer their scripts in return for an acknowledge-
ment in the source code, and some don’t even ask for that (http://
javascripts.earthweb.com/).
Likewise, many other web designers get along by using WYSIWYG editors
such as Macromedia Dreamweaver and Adobe Golive and image editors
such as Macromedia Fireworks—programs that can create many standard
JavaScript functions for you. Some respected web designers have never
programmed a line of JavaScript code; they let Dreamweaver do it.
But most web designers do learn at least the basics of JavaScript because,
sooner or later, they run into problems they cannot solve without it. A prob-
lem like this can occur: A certain page does not display properly in
Netscape 4. The solution would be to create an alternate page that does
work in Netscape 4 and use JavaScript to send Netscape 4 users to that
alternate page. For nearly every design problem like this, there is a simple
JavaScript solution.
The other problem with cutting and pasting (or relying on a WYSIWYG edi-
tor) is that browsers change, web standards evolve, and cut-and-paste
scripts as well as WYSIWYG editors tend to lag behind.
15 0732 CH11 4/24/01 11:23 AM Page 291
Hopefully, Microsoft, Netscape, and Opera will soon patch the holes in their
ECMAScript and DOM support, and Macromedia and Adobe will vastly
improve their support for these standards in Dreamweaver and Golive,
respectively. If both things happen, you might be able to spend the rest of
your life banging out advanced JavaScript functions with no clue as to
what you are doing or why it works.
If you intend to work primarily as a graphic designer and merely wish to

create simple sites for your existing clients, you can probably get by with
cutting and pasting or relying on Dreamweaver or Golive—at least for the
time being.
But if you intend to plunge into full-time web design or if you simply want
to master the craft, you will want to learn JavaScript. So let us tell you how
you can do that, and then we’ll move on to examine how JavaScript helps
web designers solve typical problems that arise in the development of any
professional site.
EDUCATING RITA ABOUT JAVASCRIPT

We’ve called JavaScript a relatively “easy-to-learn” programming lan-
guage, but it is a programming language, and teaching it is beyond the
scope of this book. In some ways, teaching it is beyond the scope of fat
books dedicated entirely to that pursuit. However, we can recommend two
books on the subject:
■ Nick Heinle’s Designing With JavaScript: Creating Dynamic Web

Pages (O’Reilly: 1997) is a wonderful, readable, detailed introduction
that any designer can understand and is chock-full of examples and
explanations of the basic terminology and theory behind JavaScript.
The book is somewhat out of date (at least as of this writing), but it
will raise your comfort level tremendously while teaching you the
basics.
15 0732 CH11 4/24/01 11:23 AM Page 292
292 HOW: The Joy of JavaScript: JavaScript Basics for Web Designers
■ JavaScript for the World Wide Web: Visual Quickstart Guide, Third
Edition, by Tom Negrino and Dori Smith (Peachpit Press: 1999) pro-
vides a series of quick exercises, complete with screenshots, that
demystify JavaScript while explaining how to perform useful func-
tions and avoid common mistakes. From plug-in testing to creating
dynamic menus or from controlling frames to baking your first
“cookie,” pretty much everything you need to know can be found
here. The scripts also are freely available at the authors’
www.javascriptworld.com site.
Speaking of free online resources, you also can learn much about
JavaScript by studying Thau’s JavaScript tutorial at Webmonkey (http://
hotwired.lycos.com/webmonkey/javascript/tutorials/tutorial1.html). Give
yourself at least two days to go through all the exercises in this five-part
tutorial. The JavaScript School at www.w3schools.com/js/ is another great
place to learn. Classic and recent JavaScript/DOM articles may be found at
http://www.javascript.about.com/.
We highly recommend that you buy these books and study these free online
tutorials. We also recommend that you take it slow, breathe deeply, and
avoid freaking yourself out over this stuff.
Don’t Panic!
As a web designer, you will not normally be expected to do advanced
JavaScript and DOM programming. Instead, your knowledge of what
JavaScript is and what it can be used for will enable you to work more
closely with team members to create engaging websites.
But don’t think you’re getting off scot-free, either.
JAVASCRIPT BASICS FOR WEB DESIGNERS

As a professional web designer, you really should be able to use JavaScript
to do simple things such as replacing meaningless URLs with text messages
as a means of extending the site’s branding. (And ducking when some vis-
itors complain about it.)
15 0732 CH11 4/24/01 11:23 AM Page 293
You should be able to create rollovers (image swaps) that help your visitors
experience the site as a responsive, interactive entity. (Yes, by hand.)
You should know how to open new browser windows (when doing so serves
a purpose), use browser detection to solve compatibility problems, and
enhance your site’s navigation through JavaScript’s ability to manipulate
simple HTML <FORM> elements.
The techniques involved are as simple to learn as they are to demonstrate.

Don’t mistake simplicity for stupidity: Some of the simple things we’re
about to show you are among the most effective ways of adding interac-
tivity to your sites.
Indeed, though we recommend learning all you can (and putting that
knowledge to use with taste and restraint), too much knowledge can some-
times lead to too much inappropriate JavaScript: scrolling text that moves
so quickly no one can read it, full-screen pop-up windows containing
rigidly designed 800 x 600 sites that look ludicrous on large monitors, or
complex, dynamic menus on general audience sites or on sites whose lack
of in-depth content is made pitifully obvious when these complex menus
end up pointing to single-paragraph pages.
While other jazz musicians blew fast and frantic, Miles Davis played very
few notes. The way he played them, when he played them, and the many
times he did not play at all, all combined to create a timeless creative
legacy. This is our highfalutin’ way of reminding you that less is more, a lit-
tle goes a long way, and slow bakin’ makes good eatin’.
So let’s look at some of these simple tasks and simple scripts. And let’s
see how ordinary web designers with no programming experience use basic
JavaScript techniques to solve everyday design and communication
problems.
15 0732 CH11 4/24/01 11:23 AM Page 294
294 HOW: The Joy of JavaScript: The Dreaded Text Rollover
THE DREADED TEXT ROLLOVER

Problem: Your client is insane about branding. In his restaurants, he brands
everything from the napkins to the silverware. He expects the same level
of branding on his site.
Solution: The JavaScript text rollover lets you brand even HTML links (see
Figure 11.1).
Figure 11.1
The status bar text
rollover in action at the
personal site of Derek
Powazek. Placing the
mouse over DESIGN FOR
COMMUNITY in the menu
bar causes the phrase
[DESIGN FOR COMMU-
NITY] to appear in the
status bar at the bottom
left of the browser. By
mastering the basic text
rollover, even beginners
can emulate at least one
Powazek design trick
(www.powazek.com).
Visit a typical site, and hold your mouse cursor over a link. You usually see
something like this:
http://www.fashionmaven.com/fashions/men/index.html
Not terribly interesting, not very informative for the average citizen, and
it certainly won’t help your brand-happy client. How much better would it
be if the visitor saw a message like this?
FASHION MAVEN fashions for men.
15 0732 CH11 4/24/01 11:23 AM Page 295
Many visitors might find this message far more useful than a bare-naked
URL. And your client would certainly dig it. Fortunately, it is easy to accom-
modate these visitors and your client with JavaScript. Text rollovers are one
of the easiest effects you can possibly create.
First, let’s look at a typical HTML link:
Explore FASHION MAVEN <a href=”/fashions/men/”>fashions for men</a>.
Notice that we’ve used an absolute link, as explained in Chapter 8, “HTML,

The Building Blocks of Life Itself.” There is no need to waste bandwidth by
including http:// or the company’s domain name in the link; both the http://
and the domain name are understood. There is also no need to waste band-
width on “index.html” because the systems administrator will have config-
ured the server to display index.html when no other document is specified.
(Some systems administrators specify welcome.html or index.htm or
default.htm instead, but the same rules apply. If default.htm is the default
document on your server, you can link to it without typing it. But we
digress.)
A visitor dragging her mouse over such a link will see the page’s URL and
nothing more:
http://www.fashionmaven.com/fashions/men/index.html
Let’s give the visitor something more informative than the page’s URL.
The Event Handler Horizon

Built into JavaScript are two powerful event handlers: onMouseOver and
onMouseOut. Event handlers enable you to create functions that take place
during an event. In this case, the event is that the visitor is dragging her
mouse cursor over a link—pretty simple stuff.
Many event handlers are built into JavaScript, but these are the two that
will help us right now. Let’s take the link just listed and make it more illu-
minating using JavaScript’s onMouseOver event handler:
Explore FASHION MAVEN <a href=”/fashions/men/” onMouseOver =”window.status=

➥’FASHION MAVEN fashions for men.’; return true;”>fashions for men</a>.
15 0732 CH11 4/24/01 11:23 AM Page 296
What is going on here?
We’ve used the onMouseOver event handler to tell the browser that some-
thing is supposed to happen when the visitor’s mouse hovers over this link.
The event handler is followed by the equal sign in the same way that links
and other standard bits of HTML use the equal sign.
As you may have guessed, window.status is JavaScript’s charming way of

referring to the status bar at the bottom of the web page. (The status bar
is the part of the browser that usually displays the bare-naked URL, gen-
erally at the lower left.) Without getting too hairy, JavaScript gives each
object in its document model an address based on the object’s position
within the document’s hierarchy, moving from the top level of the hierar-
chy down to the details: window is a top-level object; status is the object
being modified via JavaScript. (Like we said, buy the JavaScript books if you
want a better explanation, and you probably do.)
Notice that the status bar message text ‘FASHION MAVEN fashions for
men.’ is enclosed within single quotation marks. This is because the
JavaScript itself is enclosed within double quotation marks. If the text also
used double quotation marks, the browser would not know how to distin-
guish the quoted JavaScript from the quoted text.
Observe also that both the description and the phrase return true end in a
semicolon. This is basic JavaScript syntax, so get used to it. There are more
semicolons in JavaScript than in all Charles Dickens’s novels combined.
Technically, the semicolon is not required when a JavaScript statement
ends the line. So,
window.status = “some thing”
is perfectly valid JavaScript in the context of a function, a la:
<script type=”text/javascript”>
function rollover() {
window.status = “some thing” // no semicolon
}
</script>
But if you are placing two or more statements on a single line, as you would
inside an event handler attribute, you must separate the statements by
semicolons.
15 0732 CH11 4/24/01 11:23 AM Page 297
Finally, note that return true is used at the end to handle the event. It tells
the browser to follow the HTML link. Return false would tell the browser
not to follow the link, which can be useful when you don’t want to load a
new page.
Status Quo
So far, so good—now let’s make our little example even more exciting.
(Well, as exciting as this kind of stuff gets.) Let’s craft a message that shows
up in the status bar when the visitor stops hovering over the link:

➥’FASHION MAVEN fashions for men.’; return true;” onMouseOut=”window
➥status=’Welcome to FASHION MAVEN.’; return true;”>fashions for men</a>.
What have we done? (Besides further prostituting ourselves to FASHION

MAVEN, that is.)
We’ve used exactly the same syntax to replace the onMouseOver message
with a default status bar message. When the user places the mouse pointer
over the link, he’ll read “FASHION MAVEN fashions for men.” When he
releases the mouse, our insistent client will replace that link-specific brand
message with a general one: “Welcome to FASHION MAVEN.” This general
message will remain in the visitor’s status bar until he moves the mouse
over a new link. If we had not done this, “FASHION MAVEN fashions for
men” would have been “stuck” in the status bar window even after the vis-
itor removed his mouse from that link.
None of what we’ve just shown you requires any custom scripting or prepa-
ration in the <HEAD> of the HTML document. The onMouseOver and
onMouseOut event handlers are as old as the hills, and any JavaScript-
capable browser will handle this code natively. (As we’ll see later, other
JavaScript techniques require a script before the function itself.)
Well, this is fine for a single link, but coding identical onMouseOut mes-
sages for a dozen links seems like a lot of work, doesn’t it? There ought to
be a shortcut! And fortunately, there is. (Programmers are always creating
shortcuts.)
15 0732 CH11 4/24/01 11:23 AM Page 298
In the <BODY> tag of our HTML document, we can add this line of code:
<body onLoad=”window.defaultStatus=’Welcome to FASHION MAVEN.’”>
For the sake of simplicity, we’ve left out the rest of the markup you might
normally include in the <BODY> tag, such as the default background color,
text color, and so on. Of course, if you’re following W3C recommendations
and using CSS to handle your site’s stylistic elements, then your <BODY>
tag can be as simple as <BODY> with no extra junk inside it.
As you have probably deduced, onLoad is another event handler. In this

case, the event is the loading of the web page itself. When the page loads
(the event), a function must be performed. In this case, you’ve instructed
the browser that the required function is a change in the status bar mes-
sage. Thanks to your cleverness, even before the visitor hovers over a link,
the status bar at the bottom of the browser will proudly proclaim, “Wel-
come to FASHION MAVEN.” Can you feel your client’s love? We can. You
have now carried your client’s brand down to the level of the status bar.
Had you not done this, the status bar would read “Internet Zone” or some-
thing equally meaningless (as far as your client is concerned).
But wait, there’s more! Because the onLoad event handler in our <BODY>
tag is telling the browser to proclaim “Welcome to FASHION MAVEN.” by
default, we can simplify our JavaScript link as follows:

➥’FASHION MAVEN fashions for men.’; return true;” onMouseOut=”window status=’’;
➥return true;”>fashions for men</a>.
What changed? Look closely. We’ve removed the redundant text “Welcome
to FASHION MAVEN” and replaced it with Folger’s Crystals. Just kidding.
Actually, we’ve replaced it with an empty pair of single quotations, which
tell the browser to revert to the default text specified by the onLoad event
handler (“Welcome to FASHION MAVEN”), We no longer have to type “Wel-
come to FASHION MAVEN.” in the JavaScript text link itself.
That may not seem like much of an achievement. That’s because it’s not
much of an achievement. But if there are a dozen links on this page, all
requiring JavaScript text messages, we’ve saved ourselves the trouble of
typing the same onMouseOut text 12 times. We’ve also saved the viewer
the trouble of downloading those few kilobytes of redundant text.
15 0732 CH11 4/24/01 11:23 AM Page 299
Notice that it is possible to create dynamic web effects in web pages that
live on your desktop—without requiring a web server. Hooray for
JavaScript!
A Cautionary Note
Like everything you can do on the Web, modifying the default status bar
message involves trade-offs and thus requires thought. Browsers use
defaultStatus to communicate with users, letting them know if they’ve
connected or not, informing them when an object is being downloaded,
and letting the geeks in the house keep track of the actual URLs to which
your links point. Modifying defaultStatus can enhance site branding and
please your client, but it might upset some users, so don’t use JavaScript
in this way unless the benefits outweigh the drawbacks.
Kids, Try This at Home

Before we go any further, try reproducing the JavaScript effects we just
described in a simple HTML page you’ve written. Needless to say, you will
not win any innovation awards, but it might help you conquer any linger-
ing fear of programming. If you can do this simple thing, you can do other,
somewhat more complex things.
When it works on a page you’ve created, you’ll begin to feel like a web
designer. If it doesn’t work, you’ll really begin to feel like a web designer.
Then you’ll go back and fix your syntax. Speaking of which…
The fine print

Because single quotation marks are used to denote the beginning and end-
ing of text messages, what do you do if your text includes an apostrophe?
After all, in HTML, an apostrophe is exactly the same as a single quotation
mark.
What you do is “escape” the single quotation mark by inserting a backslash

character in front of it.
Lip smackin’ good! Get <a href=”/recipes/stupidcomeon.html “ onMouseOver

➥=”window.status=’Our chef\’s favorites.’; return true;” onMouseOut=”window status=’’;
➥return true;”>the recipes</a>.
15 0732 CH11 4/24/01 11:23 AM Page 300
Notice that we don’t refer to our chef’s favorites; we refer to our chef\’s
favorites. The backslash character tells the browser to treat the quotation
mark as a quotation mark, not a string terminator (meaning, not the end
of a JavaScript statement). Forgotten backslashes have caused many a web
designer her share of sleepless nights.
Return of the son of fine print

Yep, one more tip. Forget the semicolon, and you will create JavaScript
errors in many browsers, which unfortunately will not show up in many
others. That’s unfortunate because if you can’t see the error, you might not
realize it’s in there—so you may not know to fix it.
For some reason, Macs seem especially forgiving of the missing semicolon
error. Many a Mac-based web designer has uploaded a web page (or an
entire site) and gone off to smoke reefer, little realizing that he has left a
trail of JavaScript syntax errors behind him. The moral, of course, is to
check your JavaScript syntax carefully, test on multiple platforms, and
avoid smoking reefer—especially that overpriced brown stuff they’re sell-
ing uptown.
The Not-So-Fine Print

It’s worth pointing out again that some web users, including hardcore
geeks, detest this flippant toying with the sanctity of the status bar. These
users want to know which URL your link will take them to. They deeply
resent your hiding this information from them with stupid text about
FASHION MAVEN. Some might even avoid clicking the link out of paranoid
fear. (“Dude, if I can’t see the link, I don’t know where you’re taking me.”)
Thus they will never learn about FASHION MAVEN’s extensive selection of
plaids and corduroys for tall men, short men, fat men, and cadets, all at
prices 10% below what department stores usually charge.
You think we are making this up, but you haven’t read our email and
haven’t spent years watching flame wars erupt on web design mailing lists.
You think people will click links without worrying about or even noticing
these changes in the expected status bar message. Many people, of course,
won’t notice; many others will notice and not care; some will notice and
be pleased. But others will be displeased, and a few may even write letters
of complaint.
15 0732 CH11 4/24/01 11:23 AM Page 301
These people are out there, and some of them might be among your clients’
favorite customers. Thus, your infinitesimal gain in branding could be off-
set by a commensurate loss of audience. Even this small a decision is worth
considering carefully.
It’s also worth mentioning that, with the rise of HTML’s <TITLE> attribute:
<a href=”somelink.html” title=”Information about this link.”>
…there is now an easier way to enhance the information conveyed by a

link.
In IE4 (and higher), Netscape 6 (and higher), Opera 5, iCab, and Mozilla, the
<TITLE> attribute will cause a Windows-like Tool Tip or Mac OS Help bal-
loon to pop up when the user hovers over the link. (In Opera, the message
appears in the browser’s status bar, just like a JavaScript mouse-over text.)
This Tool Tip or Help balloon will contain the text you’ve written inside the
quotation marks following the word title and the equal sign. To avoid over-
whelming users with flying tool tips, there is usually a slight delay before
the Tool Tip appears. There is also no need to worry about escaped charac-
ters when writing <TITLE> attribute text:
<a href=”somelink.html” title=”It’s exciting not to worry about apostrophes, isn’t it? Gosh,
➥it’s really swell.”
Of course, if your <TITLE> text includes a double quote, the browser could
get confused:
<a href=”/” title=”We say “no!” to drugs.”>
Instead, use single quotations:

<a href=”/” title=”We say ‘no!’ to drugs.”>
Not only is this <TITLE> attribute method marginally easier to use than
JavaScript, it is also, in some ways, more logical. When a user has her eye
on a link (or a linked image), her eye does not wish to jump down to the
browser status bar. Her eye wants to say where it is. In IE4+ and Netscape
6, the <TITLE> attribute accommodates this natural behavior of the human
eye and mind because the Tool Tip or Help balloon pops up adjacent to the
link itself.
Still, we do not wish to discourage you from using status bar messages.
15 0732 CH11 4/24/01 11:23 AM Page 302
302 HOW: The Joy of JavaScript: The Ever-Popular Image Rollover
They make a handy informational and branding tool, and they work in older
browsers (like Netscape 4) that don’t support the <TITLE> attribute.
THE EVER-POPULAR IMAGE ROLLOVER

Problem: The site is pretty but feels lifeless. Visitors are encouraged to
admire but not to click and explore. The site needs a shot of GUI-like, visual
interactivity.
Solution: The JavaScript image rollover (see Figures 11.2 and 11.3).
Figure 11.2
Kaliber 10000, “The
Designer’s Lunchbox,” is a
jewel of graphic and navi-
gational design with
numerous JavaScript tricks
up its virtual sleeve. Note
the “K10k back issues”
pull-down menu at the
upper right, the code for
which is described later in
this chapter. One of K10k’s
simpler (but very effec-
tive) techniques is using
the ever-popular image
rollover to replace static
icons with animated ones.
For instance…
Figure 11.3
…dragging your mouse
cursor over the Rants and
Raves button replaces the
static dog with a GIF ani-
mation of a pooping dog.
Hey, we said they were
brilliant web designers;
we didn’t say they were
mature (www.k10k.net).
15 0732 CH11 4/24/01 11:23 AM Page 303
Let’s assume that after reading Chapter 9, “Visual Tools,” you opened Pho-
toshop and ImageReady, designed a web page comp, sliced it, and used
ImageReady to generate the JavaScript rollover. Now take those same
sliced images, open your HTML text editor of choice (Allaire Homesite,
Barebones BBEdit, or Optima-Systems PageSpinner), and, using the tech-
niques you learned in the books or online tutorials mentioned earlier in this
chapter, write yourself an image rollover by hand.
You can do it! It’s okay to prop the books open in front of you or to refer
back to Thau’s web pages. You’ll create links much like the text links we
showed in the previous example. You’ll also hand-code a preload, usually
in the <HEAD> of your document. A preload ensures that swapped images
will be downloaded to the user’s cache before the page displays. In that
way, those preloaded images are ready to leap into action the moment the
user drags her mouse over them.
Why are rollover effects so popular? We think it is because users are accus-
tomed to operating systems whose GUIs respond to their actions. Rollovers
emulate this behavior, and they indicate that an image is more than an
image—it is a dynamic trigger to an action the user can perform. Users dig
that stuff.
A Rollover Script from Project Cool

On the assumption that you haven’t bought those other books yet, haven’t
read any of the online tutorials, and still feel uncomfortable with
JavaScript, we’ll go ahead and show you another simple way to create
JavaScript image rollovers.
The following was adapted from a basic script at Project Cool. And that’s
okay. Project Cool wrote their script back in the late 1990s so web design-
ers would use it and learn from it. The future of Project Cool is doubtful
because the site’s creators left in late 1999, but this script and others like
it were still available online as of this writing (www.projectcool.com).

</script>
This script goes inside the <head></head> of an HTML document. It might

look complex if you’re unfamiliar with JavaScript, but it is really elegantly
simple.
The script begins by announcing the fact that it is a script and that its type
is text/javascript. Older browsers expected to see a <LANGUAGE> attrib-
ute with the name and, optionally, a version of the scripting language being
used (“Javascript1.2,” for instance), but this attribute has been deprecated
in favor of a more generic <MIME> type descriptor. Don’t worry if you don’t
understand what we just said; simply relax and type:
Similarly, the end of the script is announced by a </script> tag. As with

HTML and CSS, <comment> tags tell search engine spiders (and non-
JavaScript-capable browsers) to ignore everything written between . You want search engines to help web users find your content, not
your JavaScript.
Next, the Project Cool script sets a condition for running. Early versions of
JavaScript did not support image rollovers. The script wants to make sure
it is working with a browser that understands rollovers, so it tests the
browser’s receptivity to the images array object of the document model:
if (document.images)
15 0732 CH11 4/24/01 11:23 AM Page 305
The script could have accomplished the same thing by detecting for
browsers and platforms (a technique known as browser sniffing). For
instance, it could have checked for the presence of Netscape 2 and Inter-
net Explorer 3, two browsers that did not support the images array of the
document model (and hence would not be able to process this script). But
the code to check for these browsers is somewhat long compared to a sim-
ple line such as
if (document.images)
Besides, some versions of IE3 did understand image rollovers. Rather than
get tangled in browser versions, it is easier, more elegant, and more reli-
able to test for an understanding of the document images object. If the
browser does not understand (document.images), the script will be skipped.
If the required conditions are met, the script runs.
The script next declares two image conditions (Over or Out) and preloads
the required images (mb3_on-01-01.gif, mb3_off-01-01.gif, mb3_on-02-
01.gif, and mb3_off-02-01.gif):
if (document.images){
mainover = new Image; mainout = new Image;
mainover.src = “/images/menubar_over_1.gif”;
mainout.src = “/images/menubar_out_1.gif”;
storiesover = new Image; storiesout = new Image;
storiesover.src = “/images/menubar_over_2.gif”;
storiesout.src = “/images/menubar_out_2.gif”;
Over corresponds to the onMouseOver state, and off corresponds to the

default and onMouseOut state. The two images correspond to two named
JavaScript objects (main and stories).
Finally, the script declares a swapem function, which works by swapping

one image state for another:
function swapem(iname, gname) {

if(document.images){
iname.src = gname.src;
15 0732 CH11 4/24/01 11:23 AM Page 306
As we said, all of this takes place in the <HEAD> of the HTML document,
though it could just as easily live in an external JavaScript document. Like
an external style sheet as described in Chapter 10, “Style Sheets for Design-
ers” external JavaScript documents can live anywhere on the web server
and are referenced via links in the <HEAD> of each HTML page:
<script language=”JavaScript” type=”text/javascript” src=”/daily.js”></script>
For more on external JavaScripts, see “Going Global with JavaScript,” later
in this chapter.
All that remains is to call up these functions in the <BODY> of the HTML
document itself.
And here is code that does just that:

<a href =”/main.html” onMouseOver=”swapem(main, mainover); return true;”
➥onMouseOut=”swapem(main, mainout);return true;”><img name=”main”
➥src”/images/menubar_out_1.gif “ width=”200” height=”25” border=”0” alt=”Visit the
➥main page.” title=”Visit the main page.”></a>
This code should look somewhat familiar to you because it is fairly similar
to the dreaded text rollover.
Once again, here is a standard HTML link followed by two event handlers:
one for onMouseOver, the other for onMouseOut. But now, instead of
invoking a status bar message, our MouseOver and MouseOut states call
upon the swapem function declared earlier in the document. The
onMouseOver event handler declares two variables for the swapem func-
tion: a named object (in this case, main) and an appropriate image state
(mainover)—over, because this is the “MouseOver” state for the image
object. The onMouseOut event handler also declares two variables for the
swapem function: a named object (main) and an appropriate image state
(mainout)—out, because this is the “MouseOut” state for the image object.
Semicolons follow the naming of the variables and the required return true
declaration.
The image <IMG> tag that follows gives the source image a name (main),
allowing the swapem function to recognize the image as the object that is
supposed to be swapped. The remaining <SRC>, <WIDTH>, <HEIGHT>, and
<BORDER> attributes should be familiar to you from the HTML chapter. The
<ALT> and <TITLE> attributes are included so that the menu item will
15 0732 CH11 4/24/01 11:23 AM Page 307
remain accessible to those who surf with images turned off or who are
using nongraphical browsers such as Lynx. The link to /main.html will work
even if JavaScript has been turned off in the user preferences (or the
browser does not support JavaScript).
The code and the effect on the web page are much simpler than the
descriptive text you’ve just waded through.
You might ask, can JavaScript text rollovers be added to an image roll-
over like the one just described? The answer is yes, and it can be done very
easily:
<a href =”/main.html” onMouseOver=”swapem(main, mainover); window.status=’Visit the
➥main page.’; return true;” onMouseOut=”swapem(main, mainout); window.status=’’;
➥return true;”><img name=”main” src=”/images/menubar_out_1.gif” width=”200”
➥height=”25” border=”0” alt=”Visit the main page.” title=”Visit the main page.”></a>
WINDOWS ON THE WORLD

Problem: The site offers streaming video files. You, the client, or the infor-
mation architect want these files to play back inside the browser via the
QuickTime plug-in (see Chapter 12). It is easy to use the HTML <EMBED>
or <OBJECT> tags to embed a QuickTime movie in a thoughtfully designed
HTML page. But if you do this on the current page, the movie will begin
streaming even if visitors do not have the bandwidth or patience to see it.
Solution: The JavaScript pop-up window.
Opening new windows via JavaScript is a simple task, though it’s some-
what controversial. Some web users feel that everything should happen in
their existing browser window. These folks hate pop-up windows, remote
controls, and everything else that can happen outside the safe, familiar
world of their existing browser window.
Are these users right? They are right for themselves.
What does this mean? It means that pop-up windows, remotes, and other
such stunts should never be created lightly or purposelessly. (Why offend
visitors if you can avoid it?)
15 0732 CH11 4/24/01 11:23 AM Page 308
308 HOW: The Joy of JavaScript: Windows on the World
Sometimes, however, you need pop-up windows. Sometimes, nothing else

will do—as in the present example, when you wish to embed a streaming
video file in a web page but don’t want to force that streaming movie on
users who don’t care to (or can’t) view it. Pop-up windows can also be used
to provide additional information as needed (see Figure 11.4). In case of
emergency, break glass and use JavaScript to easily create new windows.
Figure 11.4
JavaScript pop-up win-
dows annoy some web
users but can be extreme-
ly functional. At TV
Guide’s site, the main
page offers a compressed
listing of all available
cable channels. Clicking
any program triggers a
pop-up window that
offers detailed informa-
tion about the selected
show. Here, for instance,
we can read about Dick
Shawn groping for laughs
as a drunken genie in The
Wizard of Baghdad. The
point is that JavaScript
allows the user to select
exactly the level of
detail needed
(www.tvguide.com).
Get Your <HEAD> Together

Before you can create a new window, you must define it in the HTML
<HEAD> of your HTML document.
Here is a typical way to do just that:
<html>
<head>
<title>Welcome to Porkchops.com!</title>
15 0732 CH11 4/24/01 11:23 AM Page 309

</script>
</head>
What are we doing? We have defined a function, given it a name (aWin-

dow), and defined its properties: It will not have a toolbar (toolbar=no), it
will be 350 pixels wide (width=350), it will stay the exact size we’ve spec-
ified (resize=no), and so on.
We have also, without even realizing it, declared a JavaScript variable—that

is, an element that can be replaced, as in the swapem example. Our vari-
able is the URL of any HTML document we would like to use in the pop-up
window.
In the HTML page, we would trigger the function like so:
<a href=”sucky_old_browser.html” onClick=”aWindow(‘porkpops.html’); return false;”>
When the event is triggered by the user’s action (clicking the link), the
named window.open function will be performed, and the appropriate HTML
page will appear in a 350 x 400 pop-up window with no status bar or menu
bar. The return false will prevent the browser from following the URL spec-
ified in the <HREF>, for backward compatibility.
As a courtesy, it’s nice to include a <CLOSE WINDOW> function in the pop-

up window itself, for the beginners in our viewing public. Porkpops.html
should include a link like this:
<a href=”#” onclick=”window.close(); return false;”>Close me!</a>
Onclick is another of those essential built-in JavaScript event handlers

you’ll come to know and love, and window.close is a built-in JavaScript
function that, as you might have guessed, closes windows. In other words,
we are telling the browser to close the window—pretty basic stuff.
15 0732 CH11 4/24/01 11:23 AM Page 310
310 HOW: The Joy of JavaScript: Avoiding the Heartbreak of Linkitis
Can we use graphics instead of HTML text to perform these functions? Oh,
yeah! In the original HTML document, we can use a fancy-pants GIF image
we’ll call openwindow.gif:
<a href=”sucky_old_browser.html” onClick=”aWindow(‘porkpops.html’); return

➥false;”><img alt=”Open new window.” src=”openwindow.gif” height=”100”
➥width=”100”></a>
And in the pop-up window we can use the dapper and elegant closeme.gif:
<a href=”#” onclick=”window.close(); return false”>)”><img alt=”Close this window.”
➥src=”closeme.gif” height=”25” width=”50”>
And that’s all there is to it.
AVOIDING THE HEARTBREAK OF LINKITIS

Problem: The client insists on a menu with dozens of choices. You know
such a menu will be ugly and confusing and will cause visitors to scroll
indefinitely (or more likely, leave). Your client “knows better.” What’s a
mother to do?
Solution: The JavaScript pull-down menu.
Slip this in your <HEAD> and smoke it:

</script>
This sets up a load_page function with a replaceable variable (which_form)

and uses the location object to swap links in and out.
Now, in the <BODY> of your HTML document, create a standard HTML pull-
down form element and use the onChange event handler to trigger new
pages in response to user actions:
<form name=”hc”>
<select name=”modules” onChange=”load_page(this.form)” size=”1”>
<option value=””>Pick a Project!
<option value=”a.html”>A List
15 0732 CH11 4/24/01 11:23 AM Page 311
<option value=”b.html”>B List

<option value=”c.html”>C List
<option value=”d.html”>D List
<option value=”e.html”>E List
<option value=”f.html”>F List
<option value=”g.html”>G List
<option value=”h.html”>H List
<option value=”i.html”>I List
<option value=”j.html”>J List
<option value=”k.html”>K List
<option value=”l.html”>L List
<option value=”m.html”>M List
<option value=”n.html”>N List
<option value=”o.html”>O List
<option value=”p.html”>P List
<option value=”q.html”>Q List
<option value=”r.html”>R List
<option value=”s.html”>S List
<option value=”t.html”>T List
</select>
</form>
This script will automatically change pages as soon as the user highlights
any item in the list. If you prefer, you can use a button or other mechanism
to actually initiate the action. You can also easily add inline CSS to add
some style to the whole sorry affair:
<select name=”modules” onChange=”load_page(this.form)” size=”1” style=”font-size:

➥10px; font-family: verdana, geneva, arial; background-color: #336; color: #ccc”>
The resulting mega-menu will look nice and take up very little space on the
page (see Figure 11.5). Compared with an endless list of standard HTML
links, the advantages of JavaScript-based navigation become obvious. To
compensate for non-JavaScript-capable browsers, you should include a
standard HTML menu somewhere on the page, but it need not be a mess if
you consolidate these HTML links using subpages:
<a href=”subpage_a-g.html”>A-G</a>
<a href=”subpage_h-n.html”>L-N</a>
etc.
15 0732 CH11 4/24/01 11:23 AM Page 312
312 HOW: The Joy of JavaScript: Browser Compensation
Figure 11.5
Add JavaScript to a stan-
dard HTML <FORM>
element, throw in a dash
of CSS for style, and you
have a tasty alternative to
the traditional navigation
menu. Instead of the mess
of links the client may
have demanded, you have
a clean, intuitive interface
requiring very little
space on the page
(www.happycog.com).
BROWSER COMPENSATION
Problem: You want to use particular technology—say, CSS—without
causing old browsers to fail.
Solution: Browser detection and redirection.
As we’ve probably boasted 100 times already throughout this book, we

publish a weekly online magazine for web designers that the gods call A
List Apart (http://www.alistapart.com/). For our 19 January 2001 edition,
we decided to create a special issue dedicated to employment problems
being experienced in the web design field at that time, due to the collapse
of many pre-IPO dot-com businesses in the last quarter of 2000.
In addition to running two articles on the subject, we were also introduc-

ing a new site feature: namely, message boards. We figured that the chance
to commiserate over business troubles would be a natural inducement to
use this new community forum.
Ordinarily, ALA’s navigational architecture employs a flattened hierarchy:

You hit the front page and are immediately presented with that week’s
content. But to highlight the special issue—to really alert our readers to the
fact that this issue was different—we decided to break with our own con-
vention and launch the issue with a splash page (see Figure 11.6).
15 0732 CH11 4/24/01 11:23 AM Page 313
Figure 11.6
This is a splash page for a
special issue of A List
Apart. Using CSS rather
than traditional HTML
tables and image slices
simplified the design and
production, reduced the
bandwidth required, and
ensured that the photo’s
color would remain con-
sistent. But this page did
not work in old, buggy
browsers. JavaScript
browser detection saved
the day (http://
www.alistapart.com/
stories/decline/).
We also decided to use CSS to lay out the page, instead of relying on the
techniques described in Chapter 10. We did this for several reasons. For one
thing, it’s leaner. Instead of an HTML table filled with dozens of image
slices, it’s three simple images, one tiny rollover image, and a few lines of
standards-friendly code:

</style>
For another thing, if we had followed the time-honored method of cutting

the comp apart in ImageReady, the colors in the photograph might not
have matched from one slice to another. And the bandwidth requirements
would have been substantially higher.
CSS enabled us to create a page that looked and worked better than tra-
15 0732 CH11 4/24/01 11:23 AM Page 314
314 HOW: The Joy of JavaScript: Browser Compensation
ditional methods allow—but there was one problem. As you’ll remember

from Chapter 10, Netscape Communicator 4 has fairly shoddy CSS support.
It does not display CSS properly and can even crash when encountering CSS
layouts.
Our referrer logs told us that 10% of our audience was using Netscape 4.
How could we offer our splash page to 90% of the audience without offer-
ing ugliness (and possible browser instability) to the other 10%?
JavaScript to the Rescue!

We solved our problem by writing a simple browser detection script and
embedding it in the <HEAD> of our HTML page:



</script>
This script defined Netscape 4 to keep an eye out for it. (We didn’t worry
about the earlier browsers because no one uses them to visit ALA.) When
a Netscape 4 user hit the splash page, he was redirected to an alternate
page via a second simple script:

</script>
15 0732 CH11 4/24/01 11:23 AM Page 315
As you can see, this script checked for a condition (browser = Netscape 4).
If that condition was met, JavaScript’s built-in window.location object
directed Netscape 4 users to main.html, the issue’s table of contents page.
The rest of the audience got to main.html by clicking the link on the splash
page. Netscape 4 users missed the splash page but they didn’t miss a drop
of content, and they didn’t realize they were missing anything. In this way
their needs were accommodated without disturbing them or any other vis-
itor to the site.
On a commercial project, we might have gone ahead and built a table-cell

version of this page for Netscape 4 users and used browser detection and
window.location to send them to that page instead.
Location, location, location

There is a drawback to using window.location. Because the redirected users
don’t realize they’ve been redirected, they bookmark the page to which
they’ve been redirected instead of the actual index page. That’s fine for
them, but when they send their friends the URL or link to the site from a
site of their own, they will be sending other users to an inner page instead
of the cover.
There is a way around that—it involves frames—but it’s a tired, messy hack,
and we don’t recommend it. If you insist on seeing how it works, visit Happy
Cog (http://www.happycog.com/), where we combine browser detection
and redirects with frames. Hopefully, by the time you read this, we will have
redesigned Happy Cog, and you won’t be able to see what we’re talking
about anyway. Never mind.
Browser detection is not always as simple as what we’ve just shown. Given
that browsers can function differently on different platforms—and because
incremental upgrades can also function differently (the 4.5 version might
choke on code the 4.6 version handles with ease)—browser detection can
get very specific and painfully complex. By a strange coincidence, we have
more to say about that very thing.
15 0732 CH11 4/24/01 11:23 AM Page 316
316 HOW: The Joy of JavaScript: Watching the Detection
WATCHING THE DETECTION

Problem: Your site absolutely requires that the user have a
plug-in installed on her system (see Chapter 12 for more about
plug-ins). Simply enough, use JavaScript plug-in detection (http://
www.javascriptworld.com/scripts/script02.08.html). But some browsers do
not understand JavaScript plug-in detection, even though they perform
many other JavaScript functions perfectly. What on earth can you do about
that?
Solution: Load o’ code—JavaScript browser and platform detection code,

that is.
Did someone say “complex browser and platform detection?” Oh, joy. An
example of that very thing follows. Specifically, it is one of Juxt Interac-
tive’s (see Figure 11.7) browser detection scripts of late 2000, written, in
part, to compensate for the fact that Juxt uses the Flash plug-in exten-
sively, and IE4.5/Mac (and earlier) did not recognize JavaScript’s plug-in
detection method—though the browser was otherwise JavaScript-capable.
Figure 11.7
The gifted designers and
programmers at Juxt
Interactive rely heavily on
the Macromedia Flash
plug-in. Juxt must be
certain its visitors have
the plug-in installed
before throwing heaps
of Flash content their way.
JavaScript plug-in detec-
tion is the answer, but
plug-in detection fails in
some browsers. Juxt’s
developers tackled this
problem by writing the
mother of all plug-in,
browser, and platform
detection scripts
(www.juxtinteractive.com).
15 0732 CH11 4/24/01 11:23 AM Page 317
If this entire chapter so far has you seriously contemplating a career as an

oil painter, we suggest you skip the next few pages, at least for now. How-
ever, we should point out that what you are about to see is not so much
complex as complete.
At first glance, the river of code you’re about to drown in looks like one
advanced function after another. In truth it is just a few functions, repeated
over and over again so that every browser version, on every possible plat-
form, can be recognized and accounted for.
The first code torrent that follows lives in a global JavaScript file called
sniffer.js. We’ll discuss global JavaScript files in a later section, “Going
Global with JavaScript,” (just as soon as we get through this section).
The second river of ‘Script lives in an HTML page called testSniffer.htm.

Let’s examine them both, shall we?
Please don’t freak. Here’s sniffer.js in all its glory:
//////////////////////////////////////////////////////
// source: juxtinteractive.com
// description: Flash 3, 4 AND 5 Detection
// Author: [email protected]
// credits: netscape communications (client sniff)
// Permission granted to reuse and distribute
// Last Modified: 10-03-00
//////////////////////////////////////////////////////
/////////////////////////////////////////
// Convert userAgent string to Lowercase
/////////////////////////////////////////
var agt=navigator.userAgent.toLowerCase();
///////////////////
// Browser Version
///////////////////
15 0732 CH11 4/24/01 11:23 AM Page 318
var is_major = parseInt(navigator.appVersion);

var is_minor = parseFloat(navigator.appVersion);
var is_ns = ((agt.indexOf(‘mozilla’)!=-1) && (agt.indexOf(‘spoofer’)==-1) &&
➥(agt.indexOf(‘compatible’) == -1) && (agt.indexOf(‘opera’)==-1) &&
➥(agt.indexOf(‘webtv’)==-1));
var is_ie = (agt.indexOf(“msie”) != -1);
////////////
// Platform
////////////
var is_win = ( (agt.indexOf(“win”)!=-1) || (agt.indexOf(“16bit”)!=-1) );

var is_win95 = ((agt.indexOf(“win95”)!=-1) || (agt.indexOf(“windows 95”)!=-1));
var is_win16 = ((agt.indexOf(“win16”)!=-1) || (agt.indexOf(“16bit”)!=-1) ||
➥(agt.indexOf(“windows 3.1”)!=-1) || (agt.indexOf(“windows 16-bit”)!=-1) );
var is_win31 = ((agt.indexOf(“windows 3.1”)!=-1) || (agt.indexOf(“win16”)!=-1) ||
➥(agt.indexOf(“windows 16-bit”)!=-1));
var is_win98 = ((agt.indexOf(“win98”)!=-1) || (agt.indexOf(“windows 98”)!=-1));
var is_winnt = ((agt.indexOf(“winnt”)!=-1) || (agt.indexOf(“windows nt”)!=-1));
var is_win32 = (is_win95 || is_winnt || is_win98 || ((is_major >= 4) && (navigator.plat-
form ➥== “Win32”)) || (agt.indexOf(“win32”)!=-1) || (agt.indexOf(“32bit”)!=-1));
var is_mac= (agt.indexOf(“mac”)!=-1);
/////////////////////////////////////
// Detect IE 4.5 on the mac
// Mucho Problemos with this browser
/////////////////////////////////////
var is_ie45mac = (is_mac && is_ie && (agt.indexOf(“msie 5.0”)==-1) &&

➥(agt.indexOf(“msie 5.5”)==-1) && (agt.indexOf(“msie 4.5”)!=-1));
//////////////////////////////////////////
// Flash 3, 4 AND 5 Detection
// Last Modified: 10-03-00
// NOT checking for enabledPlugin (buggy)
//////////////////////////////////////////
var is_flash5 = 0;
var is_flash4 = 0;
var is_flash3 = 0;
if (navigator.plugins[“Shockwave Flash”]) {
var plugin_version = 0;
var plugin_description = navigator.plugins[“Shockwave Flash”].description.split(“ “);
for (var i = 0; i < plugin_description.length; ++i) { if (isNaN(parseInt(plugin_
➥description[i])))
continue;
plugin_version = plugin_description[i];
}
15 0732 CH11 4/24/01 11:23 AM Page 319
}
if (plugin_version >= 5) {
is_flash5 = 1;
}
is_flash4 = 1;
}
is_flash3 = 1;
}
if (is_ie && is_win32) { // Check IE on windows for flash 3, 4 AND 5 using VB Script
document.write(‘<SCRIPT LANGUAGE=”VBScript”\>\n’);
document.write(‘on error resume next\n’);
document.write(‘is_flash5 = (IsObject(CreateObject(“ShockwaveFlash.ShockwaveFlash
➥.5”)))\n’);
document.write(‘is_flash4 = (IsObject(CreateObject(“ShockwaveFlash.ShockwaveFlash.
➥4”)))\n’);
document.write(‘is_flash3 = (IsObject(CreateObject(“ShockwaveFlash.ShockwaveFlash.3”
➥)))\n’);
document.write(‘<’+’/SCRIPT> \n’);
}
And now the browser and plug-in detector, as used in the HTML document:
testSniffer.htm:
<html>
<head>
<title>testSniffer - juxtinteractive.com</title>
<meta HTTP-EQUIV=”Content-Type” CONTENT=”text/html; charset=iso-8859-1”>
<SCRIPT TYPE=”text/javascript” SRC=”sniffer.js”></SCRIPT>
</head>
<BODY BGCOLOR=”#000000” TOPMARGIN=”0” LEFTMARGIN=”10” MARGINWIDTH=”10”
➥MARGINHEIGHT=”0” LINK=”#CCCC33” VLINK=”#CCCC33” ALINK=”#FFFFFF”
➥TEXT=”#999900”>
<br>
<font FACE=”Verdana” size=”2”>
//////////////////////////////////////////////////////<br>
// source: juxtinteractive.com<br>
// description: Flash 3, 4 AND 5 Detection<br>
// Author: [email protected]<br>
// credits: netscape communications (client sniff)<br>
// Permission granted to reuse and distribute<br>
// Last Modified: 10-03-00<br>
//////////////////////////////////////////////////////<br>
15 0732 CH11 4/24/01 11:23 AM Page 320
<br>
<br>
<b>Function examples</b>
<br>
(the page uses the external JS file “sniffer.js”)
<br>
<br>
<br>
<script>

</script>
<br>
<br>
</font>
</body>
</html>
Scared you, didn’t it? Scares us, too.
Don’t be alarmed. This is the province of web developers, not web design-
ers. You would not be called upon to create JavaScript this detailed your-
self. (Besides, if you ever are, you can use Juxt’s script. Note the comment:
“Permission granted to reuse and distribute,” an act of grace and kindness
that is typical of the way web designers share information with their peers.)
There are things we dislike about these torrents of code besides the fact
that they are torrents of code. Mainly we’re unhappy with the nonstandard,
old-style “extended” HTML markup. This page would not validate. As HTML,
it is not the best role model. As JavaScript, it will do ‘til the next browser
upgrade comes along.
15 0732 CH11 4/24/01 11:23 AM Page 321
Recognize that developers bash their brains out writing code like this
because browsers behave so inconsistently from version to version and
platform to platform. Be glad you’re going into web design and not web
development. Be kind to your programmers.
On the off-chance that you find this stuff enthralling or decide to switch
from design to development, you’ll find an abundance of good browser
detection information at http://webreference.com/tools/browser/
javascript.html and http://developer.netscape.com/viewsource/
krock_v5.html. Unfortunately, there is always the chance that by the time
you read this book, these pages will have moved or disappeared. If so, check
the Resources Department at http://www.webstandards.org/ for the latest
on browser detection.
GOING GLOBAL WITH JAVASCRIPT

Just as with style sheets (Chapter 10), it is possible and often desirable to
save time, hassles, and bandwidth by creating one or more global
JavaScript documents, which can then be used to control whole sections
of your site—or even the entire site.
For instance, the “My Glamorous Life” section at zeldman.com (http://

www.zeldman.com/glamorous/) is controlled by a single JavaScript docu-
ment (http://www.zeldman.com/glamorous/glam.js).
The document, in its entirety, reads as follows:
// Menubar preload. Pretty standard stuff.

function newImage(arg) {
if (document.images) {
rslt = new Image();
rslt.src = arg;
return rslt;
}
}
function changeImages() {
if (document.images && (preloadFlag == true)) {
for (var i=0; i<changeImages.arguments.length; i+=2) {
document[changeImages.arguments[i]].src = changeImages.arguments[i+1];
}
15 0732 CH11 4/24/01 11:23 AM Page 322
322 HOW: The Joy of JavaScript: Going Global with JavaScript
}
}
var preloadFlag = false;
function preloadImages() {
if (document.images) {
tocover = newImage(“../omen2/coreover.gif”);
funover = newImage(“../omen2/funover.gif”);
alaover = newImage(“../omen2/alaover.gif”);
15over = newImage(“../omen2/15over.gif”);
stealover = newImage(“../omen2/stealover.gif”);
webover = newImage(“../omen2/webover.gif”);
miscover = newImage(“../omen2/miscover.gif”);
comingover = newImage(“../glareon.gif”);
preloadFlag = true;
}
}
// Get out of some idiot’s frame.
if (top != self) { top.location = self.location; }
// Popup window, 640 x 480
function open_window6(url) {
mywin = window.open(url,”win”,’toolbar=0,location=0,directories=0,status=0,menubar=0,
➥scrollbars=0,resizable=0,width=640,height=480’);
}
// Popup window, 500 x 500
function open_window(url) {
mywin = window.open(url,”win”,’toolbar=0,location=0,directories=0,status=0,menubar=0,
➥scrollbars=0,resizable=0,width=500,height=500’);
}
Pretty “light” after all that stuff from Juxt Interactive, eh? By now it should
be obvious what this stuff means, but we’ll spell it out anyway because we
really, truly love you.
The double slashes // precede comments. The comments help the author
remember what each function is for. The double slashes tell the browser to
ignore these comments and proceed to the next function.
The menu bar preload and subsequent changeImages function are just
another way of preloading images and creating image rollovers. The images
in this case are referenced via relative URLs (../glareon.gif), as explained in
Chapter 8. It would have been smarter to use absolute URLs, but we never
claimed to be all that bright.
15 0732 CH11 4/24/01 11:23 AM Page 323
Get out of some idiot’s frame is a simple framebuster script, consisting of

just one line.
if (top != self) { top.location = self.location; }
A third-party site might link to yours. Sometimes that third-party site uses
frames. Sometimes those frames are poorly constructed. Your site might
load inside their frames instead of in its own window. This line of JavaScript
prevents that from happening. In English, what it is saying is, “The HTML
document referenced by this script should fill the browser window. If it
does, swell. If it doesn’t, get rid of any extraneous frames and fill the
browser window with our page, not some other jerk’s.” Of course JavaScript
syntax is a bit more formal than that.
The subsequent two functions are pop-up windows of varying dimensions.

They are identical except for their dimensions and their names. (The 640 x
480 window is named window6; the other is simply named window.) The
parenthetical URL (url) is a variable. If a pop-up window is needed on any
HTML page that refers to this global JavaScript document, the address of
the pop-up window will be inserted between the parentheses (popupwin-
dow.html).
How do the HTML pages make use of this global JavaScript document? Just
as with global style sheets, they do it by referring to the .js file with a link:
<script “”type=”text/javascript” src=”glam.js”></script>
The link appears inside the <HEAD> of each HTML document that requires
these scripts.
<!DOCTYPE HTML PUBLIC “-//W3C//DTD HTML 4.01 Transitional//EN”
“http://www.w3.org/TR/1999/REC-html401-19991224/loose.dtd”>
<html>
<head>
<link rel=”StyleSheet” href=”glam.css” type=”text/css” media=”screen”>
<script “”type=”text/javascript” src=”glam.js”></script>
<title>Jeffrey Zeldman Presents: My Glamorous Life</title>
</head>
<body onLoad=”preloadImages(); window.defaultStatus=’Jeffrey Zeldman Presents.
➥Entertainment, free graphics, and web design tips since 1995.’”>
15 0732 CH11 4/24/01 11:23 AM Page 324
324 HOW: The Joy of JavaScript: Learning More
Notice that the <BODY> tag includes these two onLoad functions:
preloadImages and window.defaultStatus. The first preloads the images as
referenced in glam.js. The second is our old friend, the default status bar
message—the first snippet of JavaScript we learned in this chapter. The two
are combined in one onLoad declaration and separated by a semicolon.
Simple.
LEARNING MORE
There is so much that JavaScript can do. This chapter barely hints at the
possibilities, and some methods used in this chapter could be out of date
by the time you read this book.
With the arrival of full support for ECMAScript and the DOM, the dynamic
possibilities for websites will expand exponentially. If you find, as some do,
that you take naturally to JavaScript and want to learn more about the
standardized version of JavaScript (ECMAScript) and the DOM:
■ The W3C offers the DOM at http://www.w3.org/DOM/ in all its

baffling glory.
■ WebReference’s “Doc JavaScript” (http://www.webreference.com/

js/) offers many fine articles covering ECMAScript, JavaScript, and
the DOM.
■ Peter-Paul Koch maintains a DOM mailing list (http://www.xs4all.nl/

~ppk/js/list.html).
■ The Web Standards Project maintains links to the latest ECMAScript

and DOM resources, beginning at http://www.webstandards.org/
resources.html.
And A List Apart (http://www.alistapart.com/) offers the Eisenberg DOM

series, an ongoing tutorial that includes:
■ Meet the DOM: http://www.alistapart.com/stories/dom/
■ DOM Design Tricks: http://www.alistapart.com/stories/dom2/
■ DOM Design Tricks 2: http://www.alistapart.com/stories/domtricks2/
■ DOM Design Tricks 3: http://www.alistapart.com/stories/domtricks3/
15 0732 CH11 4/24/01 11:23 AM Page 325
Whether you tackle this advanced stuff now or crawl off to recover from
reading this chapter, be proud of yourself. You have faced your fears and
at least looked at the part of web design that most designers find confus-
ing and unintuitive. This is mainly because, compared to Photoshop and
<p> paragraph tags, JavaScript is confusing and unintuitive.
But with practice and experience, it will get easier. And when browsers do
a better job of complying with ECMAScript and the W3C DOM, it will get
easier still. The programming will not be easy, but you or your development
team will take comfort in the fact that you only have to code your site one
way to work in all browsers.
There is just a little more to learn before you can consider yourself a full-
fledged (or at least a fledgling) web designer. And by a strange coincidence,
what you still don’t know is covered in the very next chapter. Let’s go for
it, shall we?
15 0732 CH11 4/24/01 11:23 AM Page 326
16 0732 CH12 4/24/01 11:24 AM Page 327
chapter 12
Beyond Text/Pictures
ON FIRST DISCOVERING THAT THE WEB IS NOT PRINT, many designers see only the
drawbacks: poor typographic resolution; a limited pool of installed user
fonts; bandwidth bugaboos; the need to compensate for browser, platform,
and hardware differences; and the awkwardness of trying to read a com-
puter screen in the bathroom.
As we start to become genuine web designers, though, most of us see more

advantages than disadvantages in the Web’s distinctive differences from
print. For example, instant worldwide distribution looks pretty darned good
after wrestling with print shops and mail houses.
The longer we work at it, the more we marvel at the Web’s ability to
provide universal access across seemingly unbridgeable gaps of technol-
ogy, nationality, economic and political systems, and physical ability or
disability.
As these barriers are crossed, the human spirit becomes less isolated, sus-
picion and intolerance begin to fade, and we learn to appreciate each
other’s differences instead of fearing them. These benefits will greatly
increase if the whole world gets to come along for the ride. They will greatly
diminish if too many humans get left behind.
16 0732 CH12 4/24/01 11:24 AM Page 328
328 HOW: Beyond Text/Pictures
This, the substance of the vision of the founders of the Web, should be
enough. But there is more. In particular, there are the two profound differ-
ences between the Web and print that we’ll discuss in this chapter:
1. The ability to develop not simply static pages, but full-fledged,

dynamic experiences
2. The visual, sonic, and interactive possibilities inherent in rich media,

whether it is delivered through emerging web standards or popular
plug-in technologies
These two unique strengths of the Web have tremendous implications for
business and for art. Each has played a huge part in popularizing the
medium. Each brims with powerful potential that designers and develop-
ers have barely begun to tap. Each also has the potential to be abused.
Figure 12.1
Nicola Stumpo’s “Destroy
Everything” is a noncom-
mercial, nonnarrative
Flash site that eats
your screen alive.
Stumpo’s emotions are
probably inexpressible
in any medium outside
Macromedia Flash
(http://www.
abnormalbehaviorchild.
com/).
16 0732 CH12 4/24/01 11:24 AM Page 329
PRELUDE TO THE AFTERNOON OF DYNAMIC

WEBSITES
In Chapter 11, “The Joy of JavaScript,” we saw how JavaScript and its big
brother, the Document Object Model (DOM), facilitate interactivity that
printed media can only dream about. In the pages that follow, we’ll look at
additional and powerful ways of making the Web more interactive.
Dynamic sites enable web users to locate information, store phone num-
bers in a shared contact database, buy holiday gifts without braving
crowded shopping centers, or view “adult” material without shame until
the baby-sitter barges in.
In this chapter, we will see how web agencies use server-side applications
to build sites that let users do things. We’ll look at where the web designer
fits in and how server-side applications help us manage immense content
sites or change text and appearance in response to user actions. We’ll also
discuss how small shops and freelancers can get in on the action even if
they don’t have casts of thousands and budgets of millions at their dis-
posal.
We’ll also see how technologies like Java can compensate for “missing
pieces” in our visitors’ browser setups or unleash full-fledged software pro-
grams that run right in the browser. And we’ll explore Java’s potential
beyond the desktop.
Figure 12.2
Here is a tranquil moment
outside the Eiffel Tower,
captured in all its
panoramic, Sensurround
glory courtesy of Apple’s
QuickTime VR—part of the
QuickTime plug-in. Print
cannot do this (http://
www.apple.com/).
16 0732 CH12 4/24/01 11:24 AM Page 330
330 HOW: Beyond Text/Pictures: The Form of Function
You Can Never Be Too Rich Media

After all that, we’ll examine emerging “multimedia” web standards that are
almost ready for prime time and take a peek and a poke at plug-in tech-
nologies that can radically enhance your sites—if used with respect for the
realities of average web users.
These technologies are not for every site, but, when appropriate, they can
enhance the web user’s experience tremendously. Used poorly, of course,
they lead to less satisfying experiences. We will explore all these tech-
nologies and consider what causes both kinds of experiences.
Knowing you as we do, we’ll start with the drier, more technical stuff
because if we saved it for later, you’d never read it.
THE FORM OF FUNCTION: DYNAMIC

TECHNOLOGIES
Think back to our earlier discussion of Perl versus JavaScript in Chapter 2,
“Designing for the Medium.” As far as the Web is concerned, Perl is most
often used in server-side transactions, such as the processing of a visitor-
submitted mail form. You might remember that a server-side technology is
one in which the computing process takes place on the web server (hence
the name) rather than the end-user’s PC. With Perl, number-crunching
tasks fall to the web server, while the visitor’s computer sits idly, waiting.
We contrasted Perl with JavaScript, whose actions take place in the

browser. With JavaScript, the end-user’s computer (the “client,” in geek
parlance) does the heavy lifting. JavaScript is a client-side technology. Nat-
urally, the dynamic technologies we’re about to consider do some work on
the client side and some on the server side. After all, the two sides are con-
tinually interacting. If the two sides, client and server, were not continu-
ally interacting, you would not have web transactions; you would just have
machines sitting around doing nothing, like Teamsters.
But though they necessarily move from one realm to the other, most of the
dynamic technologies we’re about to discuss do the bulk of their work
either on the server or on the user’s desktop. Sometimes where they work
16 0732 CH12 4/24/01 11:24 AM Page 331
is so important it becomes part of their name. For instance, as you might

guess, Server Side Includes (SSI) is a server-side technology. Mostly,
though, the names of web technologies give very little away. For instance,
would you guess, from its name alone, that PHP (originally called Personal
Home Page tools) is a server-side technology? Probably not.
Some versatile technologies work both sides of the street. Java, for
instance, is frequently used on the client side, as a downloadable applet.
But it also performs many server-side jobs. You’ll hear developers and sys-
tems administrators talk about Java servlets, which are miniature Java
applications that run the Apache server’s mod_jserv component. Or you
might host a site on Jigsaw, a W3C server that’s written entirely in the Java
language.
You don’t really have to know any of this, as long as you get the general
idea. Now let’s move on to some specifics.
Server-Side Stuff
The days of slicing Photoshop comps and hand-coding every last HTML
page are not dead—they just smell bad.
One day soon, web designers will be fully liberated from these crude pro-
duction methods. It will happen when a core group of web standards is
completely supported in browsers, enabling us to separate style from con-
tent, presentation from structure, and design from data. It hasn’t happened
yet, as any working web designer can tell you. It’s coming soon, we tell you
now. We’ll talk more about it in Chapter 13, “Never Can Say Goodbye,” so
save your questions until then.
Meanwhile, we have interim solutions that let us create web pages with-
out, well, creating web pages. Under the principles of dynamic site con-
struction, we can establish the conditions for web pages instead of building
each page individually.
The process is simple: To begin with, web designers create visual templates,
while writers, editors, and marketers create content. (Hopefully the two
teams are talking to each other so that design and content work together.)
The content is stored and indexed in vast, humming “back-end” databases,
16 0732 CH12 4/24/01 11:24 AM Page 332
and the site is launched. When visitors request data, server-side middle-
ware applications fetch the appropriate content and pour it into the
designer’s template. The result: virtual pages that can be read, used, and
bookmarked but that do not exist as conventional, self-contained HTML
documents. Oh, oh, oh, it’s magic. Let’s descend to earth and see how it
works.
Where were you in ‘82?

Ever used a search engine such as Google (www.google.com)? You type in
the name of your former high school sweetheart and hit the Google Search
button. Moments later, you’re presented with page after page of links.
From these pages you learn that your old flame is the two-term governor
of a large Midwestern state, honorary dean of a prestigious university, has
had two charities, a hospital wing, and a Ben & Jerry’s flavor named after
her, and relaxes by participating in amateur kick-boxing tournaments.
The question, of course, is why did you ever break up with her? But for our
purposes, the question is, where do these Google results pages come from?
The Google results pages are created on the fly by software that sucks
query-related entries from a huge database, determines which links are
probably most relevant, and pours the results into a preexisting HTML
template.
Who made the software? Programmers. How does data get into the data-
base? More software: specifically, a search engine spider, so named
because it crawls around the Web indexing the content and location of
individual web pages. Where does the designer fit in? The designer creates
the template that the software uses to display the results. How does the
designer do that? Let’s see.
Indiana Jones and the template of doom

As a web designer, you might be called upon to design the front end of an
application like Google, or you might work on vast content sites that rely
on similarly dynamic processing. Or you could design a site that sells things,
revealing new products in response to the visitor’s desires.
16 0732 CH12 4/24/01 11:24 AM Page 333
Paradoxically, your job will not change that much from what we’ve
described earlier in this book. You are still creating the part of the site that
the visitor sees. You design it as you would any other web project. In a way,
it’s like designing a magazine’s table of contents page. You create the mas-
ter design; someone else designs the individual issues. It’s also like design-
ing corporate letterhead in that your responsibility ends when you deliver
the approved letterhead design. You don’t have to sit and type individual
business letters. Creating website templates is as normal as those more
familiar design processes. It’s after the image pieces and HTML templates
leave your desk that the voodoo kicks in.
Precisely what happens next is up to your team’s developers—those who

write the scripts that make these dynamic transactions possible. The devel-
opers take their lead from information architects, whose job is to figure out
“user flow” through the transactional portions of the site. (Who will come
here? What will they want to do? How can we best fulfill their needs? What
can go wrong?) The very things we advised you to do when planning an
entire site, information architects do as they envision and structure the
site’s transactions.
The data can be stored in an open source MySQL database, or in similar

programs from Microsoft, Oracle, and other companies. As each visitor hits
the site and begins to take actions, the middleware that lies between the
visitor and the back-end database begins to do its thing.
It is the job of the middleware to process each request, fetch the appro-
priate document (or document fragment), and pour it into your template.
Common middleware applications include open source PHP, Allaire Cold
Fusion, and Microsoft Active Server Pages (ASP). MySQL is often found on
UNIX Apache servers, Microsoft SQL and ASP on Microsoft Windows NT
servers, and PHP can run on UNIX Apache or Windows/IIS.
Deciding on the appropriate database and middleware is not your concern.

Technology officers and network administrators solve that problem. You
aren’t expected to write code that complies with these middleware pro-
grams’ requirements either; developers do that, and we love them for it.
You can learn to write code for PHP, ASP, or Cold Fusion if you wish, and
we’ll have something to say about that in the “Doing More,” section that
follows.
16 0732 CH12 4/24/01 11:24 AM Page 334
Ordinarily, the developers and project managers will provide you with
guidelines in a document that might be called the functional spec. They will
also discuss requirements with you in one or more personal meetings—
probably more. “We can’t have frames,” they might tell you or, “we must
have frames,” could be their direction. Don’t skip these meetings and don’t
rush to argue. Talk, listen, and learn.
The work process is but a variation on what you already do. You might take
the comp no further than Photoshop; the developers will try to emulate it
in, say, Cold Fusion, and show you the result. You might ask them to revise
their code to bring the design up to your spec; they might ask you to revisit
the design to accommodate limitations in the software or particular site
requirements.
You will also write the Cascading Style Sheet (CSS) that determines colors,
type sizes, margins, leading, and so on—same as always. You might find
that some of these middleware technologies are unfortunately ill-suited to
CSS, and you might need to do some HTML table work or have it done by
your friendly neighborhood web technician.
It is sad but not surprising that some of these dynamic tools (Cold Fusion
and the like) are more suited to old-style methods of web construction
(<FONT> tags, table-based layouts, and so on) than to the newer, stan-
dards-based methods (structural markup, design via CSS). After all, these
server-side tools arose in a market driven by browser quirks and proprietary
technologies, not by universally supported web standards. As browsers
improve their support for web standards and as web designers and devel-
opers begin using these standards instead of whining about them or plead-
ing ignorance, the dynamic tools will likely improve in this regard.
Serving the project

As you might expect, database-driven sites, built with templates, are usu-
ally not the place to show off your deep Photoshop layering skills, your abil-
ity to bring complex layouts to life via frames, or your newly acquired
mastery of DHTML. Low bandwidth, large areas of flat, web-safe color, rea-
sonably sized web fonts: This is the terrain you must plow; these are the
fields you must harvest.
16 0732 CH12 4/24/01 11:24 AM Page 335
Some web designers understand this as part of the discipline of the craft
and strive to bring beauty, elegance, and utility to their simple designs.
Others rebel and might be temperamentally unsuited to this type of work.
The Web needs both kinds of designers, and there is plenty of work for both.
The need for simplicity is another reason that it’s best to do as much of the
design work as possible in CSS (as long as the middleware doesn’t choke
on it). There is little sense in asking the server to generate deeply nested
table cells when you can achieve the same result with light, clean, struc-
tural markup and a single declaration in a global style sheet. By doing the
work in CSS, you save processor cycles and bandwidth; and when it comes
time to update the design, you can do it yourself in the style sheet instead
of pestering the programmers to change their scripts.
Naturally, you will have to test to make sure that the middleware your com-
pany has chosen can handle the CSS you’ve written. You’ll also have to test
the site in multiple browsers, as described in Chapter 7, “Riding the Project
Life Cycle.” During testing, you also will want to turn off CSS in your
browsers to make sure that the resulting pages work in non-CSS browsers
(or in CSS browsers whose users have turned off CSS in their preferences).
What do we mean by “make sure the pages work with CSS turned off?" We
mean that the pages work. We don’t mean that the pages look the same
with CSS and without it. Bad clients and stupid companies expect sites to
look exactly the same in AOL 1.0 and Netscape 6. That’s impossible with-
out quadrupling the budget, and it’s also pointless. Those who turn off CSS
or use older browsers aren’t hoping for a rich visual experience. If you stick
with basic structural markup and the simple CSS techniques described in
Chapter 10, “Style Sheets for Designers,” you should be fine.
DOING MORE
Coding in PHP or ASP rarely falls within the web designer’s job description,
but after working in the field for a while, many web designers are pleased
to discover that they have a knack for these simple programming environ-
ments. If you are one of them, this knack will not go unappreciated or make
you any less marketable.
16 0732 CH12 4/24/01 11:24 AM Page 336
336 HOW: Beyond Text/Pictures: Doing More
Mini-Case Study: Waferbaby.com

Waferbaby (http://www.waferbaby.com/), Daniel Bogan’s delightful, per-
sonal site, makes smart use of PHP to facilitate dynamic content such as
the site’s “Brainstorm” section and to enable playfully user-centric design,
as seen in the site’s “Preferences” department (see Figure 12.3).
Figure 12.3
User-selectable
“Preferences” at Daniel
Bogan’s Waferbaby.
Choose a look, and the
site changes. Though
this might appear to be
the brainchild of a pro-
grammer, Bogan is actually
an animator-illustrator.
If he can do it, you can
do it (http://
www.waferbaby.com/).
In “Brainstorm,” readers respond to a provocative question on the site by

typing their answers in a form. Instantly, these answers appear on the page,
in reverse chronological order. Readers vie to outdo the wit and originality
of previous answers. The audience creates the content; personal involve-
ment and natural curiosity promote repeat visits.
In “Preferences,” visitors can modify Waferbaby’s appearance by choosing

alternate color schemes and typographic choices modeled after well-
known personal sites or create their own look and feel by editing a CSS
document right in the browser. When the reader is happy with the color
scheme and typography, it is stored as a JavaScript cookie on her hard drive.
The site will use her chosen color scheme and fonts until she decides to
change it.
16 0732 CH12 4/24/01 11:24 AM Page 337
Both “Brainstorm” and “Preferences” are made possible by a few lines of

code in PHP, a JavaScript cookie, and a MySQL database to store and fetch
the results.
Mini-Case Study: Metafilter.com

In 1999, Matt Haughey used Allaire Cold Fusion to create Metafilter (http://
www.metafilter.com/), a community site for web authors who like to write
about other people’s web content. This site will be discussed again in Chap-
ter 13. For the time being, it is worth noting how Metafilter accommodates
two levels of dynamic change: instantaneous change based on user actions
and evolutionary change based on user patterns observed over time (see
Figure 12.4).
Figure 12.4
Matt Haughey’s Metafilter
community site, a web-
based application that
responds to its members’
needs. The dark blue panel
at right, introduced in
January 2001, keeps mem-
bers posted on changes in
the way the site functions
(http://www.metafilter.com/).
As in Waferbaby’s “Brainstorm,” at Metafilter, user participation fashions

the content, generating loyalty and repeat usage. As usage patterns
emerge, Haughey responds to them by adjusting the way the site works. A
new feature is added, an old one removed. This in turn changes the way the
site is perceived and used. New usage patterns emerge, and over time, new
site-wide changes are instituted.
16 0732 CH12 4/24/01 11:24 AM Page 338
338 HOW: Beyond Text/Pictures: Doing More
Every site owner studies usage patterns and changes the site accordingly:
rotating content more frequently in the most visited sections; clarifying a
text label if one section of the site continually goes ignored; changing the
design to emphasize the least-visited section (or the most-visited); or
removing front page links to sections the public simply seems not to care
about (but keeping those sections alive to avoid link rot). The possibilities
are many.
Add dynamic, user-generated content to the mix, and the potential grows
even more interesting. On top of everything else, a psychological dynamic
begins to emerge. Is the community shaping the site, or is the site shaping
the community?
We don’t wish to imply that this whole thing goes on like a scientific exper-
iment or that the community in question serves as some kind of Petri dish
slide. As in any good community site, the owner/moderator is as involved
as any other member—but with the added ability to institute changes or
solicit suggested changes from the members. What is interesting is the way
that human dynamic behavior shapes and is shaped by basic web
dynamism. This is the power of a site that changes when you type on it.
This is interactivity print cannot match.
Any Size Kid Can Play

We’ve confined our case studies to two relatively small-scale (but influen-
tial) projects to show that dynamic interactivity is within the reach of even
the modest web shop or the lone freelancer. Large-scale projects require
teams of information architects, project managers, web designers, devel-
opers, writers, web technicians, producers, network administrators, server
consultants, marketers, advertising teams, editors, and content specialists.
If you go into full-time web design, you will likely be part of such enter-
prises. But even sites created by tiny teams can use the techniques just dis-
cussed to add web dynamism to the mix.
For instance, a dyed-in-the-wool print designer wishes to service a few of

his clients who’ve requested smallish websites: 10 to 20 pages of mostly
static content. One of these clients urgently desires the ability to post cus-
16 0732 CH12 4/24/01 11:24 AM Page 339
tomer feedback onsite in real time. Initially that seems beyond the reach
of the print designer-cum-web designer, but a few hours with a book on
PHP will change all that. The client gets his interactivity; the designer gets
a higher fee. Hopefully the consumer is also better served.
You might think all this is “too technical” for you. If you don’t believe you
will ever be able to wrap your head around server-side stuff, let us now
introduce you to Server Side Includes (SSI), the technology we mentioned
earlier in this chapter.
TAKE A WALK ON THE SERVER SIDE

As a working web designer, you might find yourself cutting and pasting the
same menu bar into page after HTML page. For instance, you might have
cut and pasted something like this into all 500 pages of your site:


<table border=”0” cellpadding=”0” cellspacing=”0”>
<tr valign=”top”>
<td width=”20%” valign=”top” align=”left” bgcolor=”#cccc00” height=”25”>
<a href=”/main.html”
onmouseover=”window.status=’Current issue. You\’re soaking in it.’; return true;”
onmouseout=”window.status=’’; return true;”>
<img name=”main” src=”/menu3/main_o.gif” valign=”top” align=”left” height=”25”
border=”0” alt=”Current issue.”></a></td>
</tr>
<tr>
<td valign=”top” align=”left” bgcolor=”#cccc00” height=”25”>
<a href=”/stories/”
onmouseover=”window.status=’Past issues.’; changeImages(‘stories’,
‘/menu3/stories_o.gif’); return true;”
onmouseout=”window.status=’’; changeImages(‘stories’, ‘/menu3/stories.gif’);
return true;”>
<img name=”stories” src=”/menu3/stories.gif” valign=”top” align=”left” height=”25”
border=”0” alt=”Past issues.”></a></td>
</tr>
<tr>
<a href=”/news.html”
onmouseover=”window.status=’Site news.’; changeImages(‘news’,
‘/menu3/news_o.gif’); return true;”
onmouseout=”window.status=’’; changeImages(‘news’, ‘/menu3/news.gif’);
return true;”>
16 0732 CH12 4/24/01 11:24 AM Page 340
340 HOW: Beyond Text/Pictures: Take a Walk on the Server Side
<img name=”news” src=”/menu3/news.gif” valign=”top” align=”left” height=”25”

border=”0” alt=”Site news.”></a></td>
</tr>
<tr>
<a href=”/join.html”
onmouseover=”window.status=’Our mailing list.’; changeImages(‘list’,
‘/menu3/list_o.gif’); return true;”
onmouseout=”window.status=’’; changeImages(‘list’, ‘/menu3/list.gif’); return
true;”>
<img name=”list” src=”/menu3/list.gif” valign=”top” align=”left” height=”25” border=”0”
alt=”Our mailing list.”></a></td>
</tr>
</table>

Ugly, isn’t it? What if you could replace that entire chunk of repugnance
with one comely line of code? Namely:

You can do it!
To do it, let’s assume that the menu mess was part of a page called
index.html.
First, cut the menu mess out of index.html, paste it into a blank document,
and save that document as menu.inc. The .inc stands for “include,” though
technically speaking, includes can have any file extension—even .html. Your
systems administrator will tell you if includes require a particular or
unusual file extension.
Now in index.html, where the menu mess used to be, type that one line:

What do these tags mean?  is an include; virtual means that what follows in quotes is a
URL pointing to the file you wish to include; and --> closes the comment
and the include.
Next, you’ll save menu.inc in an “includes” directory on your web server.

You don’t have to save it in such a directory, but it makes sense, just as it
makes sense to save GIFs in a “gifs” directory, QuickTime movies in a “quick-
16 0732 CH12 4/24/01 11:24 AM Page 341
time” or “movies” directory, and so on. As described in Chapter 7, this makes

it easier to find pieces and write appropriate file references during the site’s
creation and subsequent maintenance. If for some reason you prefer to
save your SSI files in a directory called “rosebud,” the reference would read:

Now simply use that line of code in every HTML document where you for-
merly had to cut and paste a heap of menu bar markup. Then upload your
HTML pages to the web server.
Some folks use a different file extension, such as .shtml or .shtm, if their
HTML file contains an include, and some servers require this. But if you can
stick to the .html file extension, you’ll avoid confusion and heartache down
the road.
Why confusion and heartache? We knew you were going to ask. For one
thing, imagine that your static .html pages have been bookmarked by vis-
itors and search engines. You then start changing your file extensions. All
of a sudden, your internal and external links are broken, your visitors are
confused, and the search engines that ranked you so highly are pointing to
nonexistent pages.
Are You Being Served?

You’ve replaced redundant markup with neat, clean includes. What’s the
next step? There probably isn’t one. Most web servers natively support SSIs.
If it doesn’t work right away, you might need to contact the company host-
ing your site (or the network administrator if your company hosts its own
sites) and ask that the configuration file be changed to permit SSIs. Unless
the hosting company hires trained monkeys as tech support, complying
with your request will take two minutes.
Of course, if you are sane, you will have made this phone call before chang-
ing all your HTML pages. Or you will have created a test HTML page,
uploaded it, and confirmed with your own eyes and mouse cursor that it
works.
16 0732 CH12 4/24/01 11:24 AM Page 342
342 HOW: Beyond Text/Pictures: Take a Walk on the Server Side
More than one SSI can be put to use on each page. You can replace the
“header,” the “footer,” or just about any piece of the puzzle. Using SSI, you
can replace all or nearly all of the dull, repetitive junk that holds web pages
together.
In turn, you can begin viewing HTML pages as content containers rather
than tortuous masterpieces of visually oriented markup—because content
containers are exactly what they are and were always intended to be. This
might not be the true separation of style from content, but it will do until
the real thing comes along.
SSI can do many things besides what we’ve outlined here. It can insert
appropriate text, HTML, or CSS based on the user’s browser. It can indicate
when the page was last updated (),
give the current date and time, and do other funky tricks.
And, as we’ve said, SSI is the low end. Imagine the possibilities if you begin
to work with more advanced server-side technologies.
Advantages of SSI
If a site changes—or perhaps we should say when a site changes (for
instance, when a new section must be added to the menu bar)—the power
of SSI is revealed. What was true for CSS is just as true for SSI: It is easier
to edit a single document (menu.inc) than it is to change hundreds or
thousands.
Hopefully, your client is not about to wantonly add new sections to the
menu or demand changes to the appearance of the menu after the site is
nearly built. In a perfect world, you have followed the suggestions in Chap-
ter 7, and the client has signed off (and paid part of your fee) at each stage
of completion. Therefore, the client has a vested interest in following
through with the plan he committed to and paid for and has no vested
interest in pulling last-minute changes to prove that he is the dominant
monkey in this rainforest.
But clients are clients, and change happens. SSI is a simple way of pro-
tecting yourself from hours of tedious replacement tasks.
16 0732 CH12 4/24/01 11:24 AM Page 343
Disadvantages of SSI
Being a server-side technology, SSI eats up processing power from the
server. Every time a visitor hits a page containing a replaced SSI element,
the markup describing that element must be fetched from the server—like
a stick in the jaws of a panting mastiff.
If you’re building a professionally hosted site with plenty of server power

in reserve, such demands on the server are no problem. If you’re hosting
the site on your home computer and connected to your home cable modem,
there could be a problem. Given sufficient traffic, the toll on your PC might
be noticeable, and the site might “slow down” for your visitors during times
of peak traffic. On the other hand, if you’re hosting an extremely popular
site on your home computer, maybe it’s time to upgrade your server.
If you are interested in server-side technologies, Jeff Veen’s The Art & Sci-
ence of Web Design (New Riders: 2001) discusses the subject in more
detail—and using better words and stuff. If you are uninterested in server-
side technologies to the point of anxiety, you’ll be happy to know that
we’ve finished discussing them.
Now let’s look at a technology you will frequently encounter in your career
but will never even contemplate programming yourself. Let’s talk about
Java. First of all, what the heck is it?
COOKIN’ WITH JAVA

Java is an object-oriented programming language developed by Sun
Microsystems (http://www.sun.com/) primarily for the Web. And just what,
you ask, is an object-oriented programming language? An object-oriented
language is one that reuses software objects the same way you might re-
use custom shapes you’d created in Adobe Illustrator or a sales executive
might reuse chunks of boilerplate text about “tremendous synergies,
should our two companies work together.”
16 0732 CH12 4/24/01 11:24 AM Page 344
344 HOW: Beyond Text/Pictures: Cookin’ with Java
In Illustrator, you can recombine basic button shapes, spirals, or complex

outlines to create new artwork from predesigned fragments. Similarly, a
Java programmer can combine entire libraries worth of coded objects to
build new programs from existing parts. Reusing graphic elements makes
you faster and more productive; reusing code objects does the same thing
for Java programmers.
Reusable parts: that’s the idea. Sun’s programmers called these parts
objects. Sun didn’t invent this idea. Windows, Mac OS, and UNIX also reuse
code objects (Windows DLLs, anyone?). But in operating systems like Win-
dows, Mac OS, and UNIX, these reusable parts are immediately compiled
down to machine code. In Java, they are compiled to an intermediary for-
mat called “bytecode,” which is then interpreted by a Java Virtual Machine,
about which we’ll have more to say in just a moment.
As mentioned earlier in this chapter, Java can be used to create full-scale

programs (applications), miniature programs that download quickly when
needed (applets), or server-side servlets. Servlets are full-fledged but small
application fragments that run in the context of the server—as Photoshop
plug-ins run in the context of Photoshop.
Ghost in the Virtual Machine

But there’s a catch. Just as Windows programs require a Windows envi-
ronment and Mac programs are designed for Macs, Java programs must run
in a Java environment.
Does this mean that you have to go out and buy a Java computer? No, it
simply means that Java programs are designed to run in Java-capable web
browsers (Netscape Navigator, Microsoft Internet Explorer), Java-capable
web-enabled devices, or special Java devices (such as Java-powered digi-
tal television-top boxes and remotes). They do this by means of Java Vir-
tual Machines, which we promise, really truly promise, we will describe in
just a moment.
Netscape was the first browser to support Java, and the point of the
Sun/Netscape partnership, as explained in Chapter 2, was to smash Win-
dows hegemony while getting Java onto as many platforms as possible, by
way of the browser. They succeeded at getting Java onto as many platforms
as possible. One out of two ain’t bad.
16 0732 CH12 4/24/01 11:24 AM Page 345
Today most browsers and computer operating systems support Java. It gets
a bit more complicated when the browser or OS maker offers an “improved”
Java environment that Sun does not consider truly Java-compatible, but
we’ll get to that later. Java-capable browsers might run on any computing
platform (Windows, Mac OS, Linux, UNIX, or BeOS) as long as the browser
manufacturer supports that platform.
What makes all of this work? The Java Virtual Machine does. You might
think of the Virtual Machine as a streamlined computer operating system
(OS) running inside another computer OS—a Java computer running inside
Windows, for example. Or you might think of it as an interpreter, turning
spoken words into sign language for the hearing-impaired.
This Virtual Machine is sometimes included with the browser. Early versions
of Netscape included a Virtual Machine customized for each OS. This added
significantly to the download time but ensured that users would have the
then-new Java technology at their disposal.
In other cases, the Virtual Machine is built into the operating system. For
instance, Apple Macintosh OS9 includes “Mac OS Runtime for Java,” a Java
Virtual Machine whose sole purpose is to run Java programs on the Mac.
If you install IE5 Macintosh Edition on a pre-OS 9 Mac, you might get Java
errors because IE5/Mac expects a more recent Virtual Machine than the
one on your system. You can correct this problem by upgrading to OS9 or
by downloading a more recent version of Mac OS Runtime for Java from
http://developer.apple.com/java/classic.html. The program is free.
As you can see, the tantalizing potential of Java lies in its ability to work
in any operating system equipped with a Java Virtual machine—in other
words, theoretically at least, to run on any operating system. Practically
speaking, developers could build a word processor or a full-blown office
suite that runs in any Java-capable web browser and on any operating sys-
tem with a Virtual Machine. Of course, companies that make word proces-
sors and full-blown office suites might not like that idea. They might dislike
it so much that they would end up building their own web browser and tak-
ing over the market…not that we’re mentioning any names. There is, in fact,
a Java word processor (indeed, there is an entire Java office suite), and we
hear it works quite well.
16 0732 CH12 4/24/01 11:24 AM Page 346
Where the web designer fits in

As a web designer, you might be called upon to embed a Java applet in an
HTML page. (Again: An applet is a self-contained piece of code that runs
within a Java-capable browser, as Photoshop plug-ins run within Photo-
shop.) This is simply a matter of using the HTML <OBJECT> or <APPLET>
tag or another very basic HTML tag—no problem at all. At other times, you
might use Java to compensate for a missing plug-in on a visitor’s system.
For instance, the IpixViewer plug-in, like Apple’s QuickTime VR (see the sec-
tion, "Turn on, Tune in, Plug-in" later in this chapter), enables visitors to
explore 360º panoramic views of any location that can be photographed.
It’s an extraordinary plug-in that does a remarkable job. But not many peo-
ple know about this plug-in, so not many have downloaded it. Therefore
you might feel that IpixViewer content cannot be used on your site. Not to
worry! The missing plug-in can be replaced by a Java applet and compiled
down to native, platform-specific code via the Java Virtual Machine:
<applet name=”IpixViewer” code=”IpixViewer.class” archive=”IpixViewer.jar” height=”210”

➥width=”280”>
<param name=”URL” value=”zabptcaj.ipx”>
<param name=”Spin” value=”on”>
</applet>
If the HTML just listed looks odd to you, don’t sweat it. Your Java developer
will tell you what needs to be included on the page. Your job will be to
insert it, test it, and verify things such as height and width. (Is the result-
ing image in fact 210 pixels high? Does it look right? If not, change the
numbers and try again.) By the way, this same technique works for other
multimedia content, such as Flash. If the visitor lacks the Flash plug-in, a
Java applet can display the Flash content. Your developers will create the
applet and the complex code that determines whether or not the applet is
needed on each visitor’s system. Your job is simply to plug in some HTML
and test.
The other reason you need to know about Java is that in spite of its utopian
aims and utilitarian benefits, Java can sometimes be problematic. And as a
user-oriented web designer, you need to be aware of that.
16 0732 CH12 4/24/01 11:24 AM Page 347
Java Woes
We can do this two ways: the short, brutal version or the long, boring, polit-
ically correct version.
Here’s the short, brutal version: From a user experience perspective, Java
often sucks. It can be as unstable as Norman Bates, drain resources like
Australians drain beer steins, and crash more frequently than a drunk dri-
ver’s Pinto.
For those who expect us to be fair, a long, carefully guarded, politically cor-
rect version follows. Feel free to skip it unless you are an attorney for Sun
Microsystems. In which case, we meant to say that Java is the best thing
since the Magna Carta.
Java Woes: The Politically Correct Version

At times, companies have created their own Java Virtual Machines that dif-
fer subtly from Sun’s. Sun does not like that, and you can understand why.
Java is not open source; it is a protected product. Differing Virtual
Machines can sometimes prevent Java from fulfilling its promise. This has
led some developers to avoid using Java. As a web designer, you will want
to stay aware of these issues if there’s a possibility of their affecting your
site and your users.
Java can also sometimes drain the computer’s memory resources because
the user is essentially running a second operating system (Java) within his
existing OS. Not to mention the fact that the user is likely running a Java-
based application on an unstable web browser with all its memory-hog-
ging plug-ins, on top of any other software programs he might have
running in the background, and on top of a possibly unstable base operat-
ing system such as an older version of Mac OS, which can be wonderful but
not entirely stable.
The older the computer and the less memory at its disposal, the greater the
possibility of woe. Attention, Sun attorneys: We do not wish to overstate
these issues. All that is usually required is for the user to increase the
amount of memory allotted to the browser. Unfortunately, most web users
don’t realize this, so they don’t do it. Result: instability.
16 0732 CH12 4/24/01 11:24 AM Page 348
The memory problem is not a Java problem per se; plug-ins like Flash and
Shockwave also work better if the user increases his browser’s memory par-
tition. Fortunately, during the installation process, Shockwave and Flash
alert users to the issue and offer to increase the browser’s memory auto-
matically if the user clicks the OK button. Java does not do this because
Java is typically preinstalled on the user’s machine when it arrives from the
factory.
Given that browser makers know most users are going to encounter Java
and are going to install and run plug-ins, why don’t they increase the
default memory partition of their browsers? In a word: competition. The
browser makers want to prove that their product uses less memory than
the competitor’s, so the browser installs itself with the lowest memory
allotment possible. It will operate under those conditions just fine as long
as users rarely venture beyond all-text websites. Most users do venture far
beyond, whether knowingly or not. So most users are practically guaran-
teed to encounter browser instability on sites that use Java or plug-ins or
even large, memory-draining background images.
Though Java tends to work well in Windows and UNIX, it’s a mixed bag in
Mac OS. Even on top-of-the-line G4 Macs with 1.5GB of installed RAM, T3
connections, and system buses capable of transferring over 1GB of data per
second, Java can sour.
These same Macs can rotate a 40MB Photoshop image faster than Google
can track down your ex-girlfriend. At speeds exceeding 5.5 gigaflops, they
can outperform Pentiums with twice the rated clock speed. But a stupid
“rippling water” Java applet on a personal site at Geocities can take down
these mighty Macs. Java is cross-platform but not always reliably so.
Attention, lawyers: We do not wish to overstate these issues.
Then of course, Java does not work at all in text-based browsers such as
Lynx, nor will it function in older browsers such as Internet Explorer 2. And
users of even the newest browsers might “turn off” Java in their prefer-
ences, thus defeating the development team’s efforts to use Java on the
site.
16 0732 CH12 4/24/01 11:24 AM Page 349
This is not a Java problem per se. Users can also turn off JavaScript and
style sheets. They can refuse to install plug-ins, tell the browser to use
“their” background colors instead of yours, and in every other way imagi-
nable assert their right to see the Web as they wish to see it, thus turning
your beautiful site into a sea of sewage that strangely pleases them.
The workaround, as always, is to provide alternatives. Simple HTML menus

and alternative content go a long way toward keeping sites accessible, no
matter what technologies are intended for their use under optimal condi-
tions. We do not wish to understate this issue. We wish to strongly empha-
size it. Make accessibility part of the plan at all times.
Java Joys
Despite hiccups, Java is cross-platform, and it does many things very well,
such as “stepping in” to replace missing plug-ins. For instance, as just
described, Flash files can be run as Java applications in Netscape Naviga-
tor if the user does not have the Flash plug-in. That is fairly remarkable. It
is handled by Flash itself. When saving the file, Flash generates code that
will call upon a built-in Java action if the plug-in is not detected in the
user’s browser.
Beyond all that, Java applets and Java servlets (smaller, more stable mini-
applications of Java that run on the server) can be used to help create
dynamic, database-driven websites. Java is ideally suited for sophisticated
tasks that take place under the hood. Because Java works cross-platform
and cross-browser (despite problems just mentioned), it might be prefer-
able to use Java for complex tasks, rather than relying on proprietary, plat-
form- and browser-specific technologies such as VBScript and ActiveX.
Java seems less valuable to us when it is used to create dynamic menus or

to trigger the rotation of ad banners. In both cases, JavaScript/ECMAScript
is a lighter, more stable choice that is also a web standard, tends to use
fewer computing resources, and works better across platforms.
16 0732 CH12 4/24/01 11:24 AM Page 350
350 HOW: Beyond Text/Pictures: Rich Media
Seeing as we’ve mentioned Java and JavaScript in the same paragraph, we

might as well restate that the two technologies should not be confused, in
spite of their similar names. JavaScript is a complex but interpreted pro-
gramming language that works in web browsers. Java is a full-fledged,
object-oriented programming environment that can drive entire devices or
can be used to build complete applications. Nearly all web designers work
with JavaScript, whether on the programming level or simply via cut-and-
paste. No sane web designer attempts to program in Java. Even insane web
designers avoid it.
The true power of Java is now being manifested beyond the browser.
Instead of web surfing, consumers are channel surfing via Java-powered
TV devices (www-us.semiconductors.com/news/content/file_501.html).
Java and Linux are now creating Internet appliances that require no
understanding of Java or Linux (http://www.linuxdevices.com/news/
NS5323294840.html). Java is finding its way into Personal Digital Assis-
tants (PDAs), cell phones, and even server-side technologies (http://
www.alistapart.com/stories/beyond/2.html). Keep your eye on Java as your
career unfolds, and use it judiciously as your sites evolve.
RICH MEDIA: EXPLODING THE “PAGE”

We say web “pages” because our minds cannot let go of the publishing
model we grew up with. But rich web media give the lie to the “page”
metaphor. These pages are not pages. This is not a pipe. This is not my beau-
tiful wife.
Let’s see how standard technologies and popular plug-ins push the Web
way beyond the cosmologies of print design. We’ll start with some web
standards you might or might not know about.
Virtual Reality Modeling Language (VRML)

VRML, though nearly dead from disuse, is the standard language for the
animation and 3D modeling of geometric shapes. It allows 3D scenes to be
viewed and interactively manipulated on the Web. Using a special VRML
browser, the web user can connect to an online VRML site, choose a 3D
16 0732 CH12 4/24/01 11:24 AM Page 351
environment to explore, and cruise around the spooky “3D world.” It is also
possible to zoom in and out and to interact with the 3D environment in
various ways. The Netscape Live3D VRML browser (built into Netscape 3)
was the first to support the VRML 1.0 standard.
Think video game. Think cheesy, super-low-grade video game. Laura Croft
it’s not. It’s more like Pacman 3D. Think wireframe and black backgrounds.
Besides being fairly crude, VRML is not a technology that lends itself to
accessible alternatives. A GIF image might be described via <ALT> and
<TITLE> text for the benefit of web users with visual disabilities. But you
are either navigating a 3D environment, or you’re not. <ALT> text just
won’t cut it: “If you could see and if you could physically manipulate a
mouse cursor, you might enter a crude simulation of a living room and ‘pick
up’ an illustration representing a pencil.” Thanks for sharing.
VRML is fascinating but has few immediately apparent commercial bene-

fits. Nor is it particularly dazzling in today’s world of Flash 5, DOM-based
interactivity, and improved monitor and color resolutions. Perhaps for these
reasons, the technology has never caught on the way that JavaScript, for
example, caught on. Web users have a tough enough time finding what
they want on most websites without adding primitive 3D effects to the mix.
Of course, VRML was never about “web users finding what they want,” and
you might feel we’ve just slapped a straw man. But have we?
Web-using veterans might recall a similar 3D experiment called Hot Sauce

that was created by Apple Computer in the mid-1990s. Hot Sauce turned
text-based directories into virtual 3D environments containing (you
guessed it) text—text that floated in fake 3D space. To move from one block
of text to the next, you eased your mouse up and down your desk.
Instead of navigating Yahoo.com the conventional way and finding what

you wanted in under 30 seconds, with Hot Sauce you could spend hours
painfully navigating a 3D version of the Yahoo directory. This was not most
people’s idea of fun, and the technology soon petered out. Scientists do
what they can; marketers do what sells. Hot Sauce did not sell, and neither
did VRML because after you muttered, “Cool,” there was little else to
be said. Angry VRML and Hot Sauce fans, please send your protests to
[email protected].
16 0732 CH12 4/24/01 11:24 AM Page 352
Now that you know what VRML is, you probably don’t need to know much
more about it. If you’re curious, more information is available at The Web
Developer’s Virtual Library:
http://wdvl.internet.com/Authoring/VRML/
SVG and SMIL

In the absence of finalized multimedia standards for the Web, plug-ins
were developed that enabled websites to offer streaming video, animated
vector graphics, music tracks, and the like. We are about to look at those
very plug-ins. But first, let us pause to consider a recent development.
Over the past couple of years, W3C recommendations have emerged to

suggest standardized ways of doing what proprietary plug-ins already do
so well. One of these is SMIL, the W3C recommendation for multimedia;
the other is SVG, intended to deliver vector graphics such as those already
used in Flash (but with some essential differences from Flash).
What’s up with these two new standards, and why do they matter?
SMIL (through your fear and sorrow)

SMIL (http://www.w3.org/AudioVideo/) stands for Synchronized Multime-
dia Integration Language and is pronounced, “smile.” Isn’t that cute? Oh,
shut up.
SMIL is an easy-to-learn, HTML-like language for creating “TV-like multi-

media presentations such as training courses on the Web,” according to the
W3C. The current SMIL recommendation is 1.0, and you can read all about
it at the W3C address just cited and at another one we’ll mention later. This
is our way of avoiding adding another 50 pages to this book.
Aside from the fact that three Internet heavies (Real, Apple, and Adobe) are
throwing their weight behind SMIL, why should you care about any of this?
Let’s see.
Harnessing media, helping users

SMIL packs accessibility features (http://www.w3.org/TR/SMIL-access/),
including alternative text content that can be made available to Braille
readers. Such content will also enable search engines to index multimedia
web content authored in SMIL.
16 0732 CH12 4/24/01 11:24 AM Page 353
In English: slap a QuickTime video on your site and search engines such as
Google or Altavista could care less. But add a carefully authored SMIL pres-
entation to your site, and speeches made by the characters in your video
could show up in Google and Altavista’s search results.
The educational implications are enormous. A student researching Ham-

let’s soliloquy could find a SMIL-authored video of Sir Laurence Olivier per-
forming it. The Web’s potential as the world’s library could suddenly
become much richer.
The commercial implications ain’t bad, either. A buyer searching for widg-
ets could find your client’s digitized promotional video on the subject.
Existing multimedia formats obviously do not offer these advantages.
Lest you think SMIL is a completely wacky new technology, it is, in fact,
simply a markup language that works with existing technologies like
QuickTime and Real digital video and audio. What SMIL does is bring the
traditional benefits of the Web (searching, finding, bookmarking) to non-
text content. That is profound.
More reasons to SMIL

Other cool things you can do with SMIL:
1. With a single link, you can deliver audio to dialup users and video to
broadband users. None of that “click here for audio, click here for
video” junk.
2. Deliver different language versions of clips depending on a user’s

system-language setting.
3. Use back-end technologies to deliver multimedia content on the fly.
No need for expensive, proprietary programs with steep learning
curves. (SVG delivers similar benefits.)
…All with a few simple tags.
Author! Author!
Among the currently available Web tools and plug-ins that support SMIL
are Apple QuickTime 4.1 (http://www.apple.com/quicktime/) and the
unfortunately named RealSlideshow authoring tool by the makers of
16 0732 CH12 4/24/01 11:24 AM Page 354
the RealPlayer (http://proforma.real.com/rn/tools/slideshow/index.html?).

Adobe is presently developing a SMIL extension for its GoLive WYSIWYG
tool, which should simplify the creation of SMIL content and might help
accelerate the standard’s adoption.
RealSystem’s support for SMIL has been solid since 1998. Given the num-
ber of RealPlayers out there, SMIL can already reach almost as many web
users as Flash does. Not that SMIL and Flash are enemies. SMIL is often
used to integrate Flash content into the QuickTime and RealPlayers, and
Flash 5 exports SMIL for use in RealSystem.
SMILsoftware’s Flution 1.5 (http://www.smilsoftware.com/) for Windows

can streamline the SMIL creation process. Tom Wlodkowksy’s free Media
Access Generator (MAGpie) for Windows (http://main.wgbh.org/wgbh/
pages/ncam/webaccess/magpie/) adds accessibility features such as
closed-captioning to SMIL. For a more detailed description of the goals of
the SMIL language, see the W3C Activity Statement (http://www.w3.org/
AudioVideo/Activity.html) on Synchronized Multimedia. For practical
advice on putting SMIL to work, see Jim Heid’s old-but-good tutorial at
Macworld, SMIL: Markup for Multimedia (http://macworld.zdnet.com/
2000/02/create/markupmultimedia.html).
SVG for You and Me

SVG (Scalable Vector Graphics) is a W3C standard in progress. As of this
writing, the W3C describes its initial SVG activities as “currently nearing
completion” (http://www.w3.org/Graphics/SVG/Overview.htm8). Though
SVG produces vector graphics, it is a markup language. In fact, it is an
application of XML, the super-meta-markup language we’ve mentioned
throughout this book.
Like Flash vector graphics, SVG vector graphics can fill an entire screen with
artwork while using very little bandwidth. Also like Flash, SVG can be ani-
mated via scripting. You’ll find examples of this at Adobe’s SVG site, which
we’ll discuss in a moment (see Figures 12.5 and 12.6).
16 0732 CH12 4/24/01 11:24 AM Page 355
Figure 12.5
The Battlebots logo in
SVG. At the user’s discre-
tion, the image can be
enlarged again and again.
Figure 12.6
Vector artwork maintains
quality at the highest
magnifications while keep-
ing bandwidth expenditure
at a minimum (http://
www.adobe.com/).
No matter its graphic appearance, SVG remains text. To understand the

implications of that fact, let’s contrast SVG with our present production
techniques. We’ll use an example that’s close to every designer’s heart: the
client’s logo.
16 0732 CH12 4/24/01 11:24 AM Page 356
Romancing the logo

For the purposes of this little exercise, we’ll assume that the client’s logo
involves letterforms rather than nonverbal swooshes or swirls. We’ll further
assume that you’re developing the logo in Adobe Illustrator and that you
have not yet converted your text to outlines. After it becomes outlines, it
ceases to be text, thus losing the SVG benefit we’re about to explore.
First, the traditional methods:
If you export your client’s logo from Illustrator to Photoshop and embed it
on a web page as a GIF image, search engines will not index it because it
is not text. You can work around that limitation by adding <ALT> text to
your image tag, but not all search engines index all <ALT> text.
If you create that same logo in Flash, it can spin and whirl and glow, but
search engines will not index it because it is not text. Flash 5 has added
some accessibility features, allowing you, for instance, to include <ALT>
text for a Flash file, but this is global text, not image-specific text, and we
already talked about the limitations of <ALT> tags as a guarantor of search
engine placement.
Now, the SVG method:
Take that same Illustrator logo and export it as SVG, using Illustrator’s
built-in support for that web standard. The resulting logo looks great,
smells fresh, and it remains text. That means search engines can index it.
Your client’s logo no longer blushes like a maiden when the search engine
comes courting. From every page of the site, the text-based logo calls out
to the search engine, and the search engine rewards it with the Web’s
greatest mark of love: a high ranking.
To the eye, the logo is a logo; to the search engine, it is a word. If the word
“Widgets” appears at the top of every page of the site, that site will rank
high when users search for widgets. When the client cries, “Make the logo
bigger,” you can answer: “We’ve made it number one.” By contrast, under
the old methods, when a GIF image of the word “Widgets” appears at the
top of every page of the site, it is unlikely to seduce the search engines.
16 0732 CH12 4/24/01 11:24 AM Page 357
Because the SVG-formatted Widgets logo is a word that looks like a logo,
users can also copy and paste it into a text document. It will lose its SVG
formatting when users do this, but your client’s name will remain intact.
Your client will like that. And who knows? A year from now, it might not
lose its formatting when pasted into a popular word processor, print lay-
out program, or email message.
In fact that is one of the promises of SVG for graphic designers: that we
will be able to use the same SVG image file in our print work and our web
work—from Illustrator to Quark to the website, as easy as drag and drop.
(Yes, you can also create SVG illustrations by hand-coding them—after all,
SVG is really XML—but we doubt many designers will want to do that. We
sure don’t.)
Will SVG replace Flash? Not likely and certainly not any time soon. Will SVG
evolve into a useful tool for creating scriptable vector graphics? We think
it will.
Sounds dandy, but will it work?

SVG support is coming online slowly. A plug-in from Adobe (http://
www.adobe.com/svg/main.html) supports SVG in all web browsers, though
not equally well. The first version of the Adobe plug-in relied on Netscape-
proprietary plug-in detection that was not supported in Internet Explorer
for Macintosh. Users of IE5/Mac could not see SVG graphics at all with that
plug-in version.
As of this writing, a newer Adobe SVG plug-in has greatly improved its sup-
port for non-Netscape browsers, though Internet Explorer support for Macs
is limited to nonscripted SVG only. In other words, IE/Mac users can
see SVG graphics on the Web but cannot see dynamic (animated) SVG
graphics.
Still, things are looking up for SVG. You might find it odd that it takes a
proprietary plug-in to support an open standard, but such is the state of
the Web. After the SVG standard is finalized, we suspect that browser mak-
ers will begin investigating ways to support it.
16 0732 CH12 4/24/01 11:24 AM Page 358
358 HOW: Beyond Text/Pictures: Turn on, Tune in, Plug-in
Promises, Promises
While SMIL is well-supported via plug-ins, SVG is still a work in progress—
a promising work in progress, an exciting work in progress, but a work in
progress. It will be widely adopted when it is further along in development
and can be natively supported in browsers.
Do these tools, SMIL and SVG, pose an immediate threat to Macromedia

Flash? They certainly do not. In fact, we don’t see them as anti-Flash tech-
nologies at all (though some might view them that way).
While SMIL is expanding and SVG is still taking shape, now would be a good
time to download Adobe’s SVG plug-in and explore SMIL presentations and
tutorials. Over the coming months, you will want to remain open-minded
about these emerging standards and keep your eye on their evolution.
Before we know it, they will likely be part of every web designer’s tool kit.
But in the meantime, you have a job to do. So let us turn our gaze to the
Web’s de facto multimedia “standards” (which are not, technically speak-
ing, web standards at all). Let’s consider the proprietary, often-maligned,
sometimes-adored, widely used plug-ins that already bring rich multime-
dia experiences to hundreds of thousands of sites and hundreds of millions
of web users.
TURN ON, TUNE IN, PLUG-IN

Plug-ins are the chief means by which web designers currently add sound
and motion to the Web, and web users employ them to extend the capa-
bility of their browsers, allowing them to see and hear these sound and
motion effects. Browser can’t play music? Pop in a plug-in. Browser can’t
show vector graphics? Pop in a plug-in. Browser can’t show 360-
degree panoramic views of the client’s flagship $599 running shoe? Pop in
a plug-in.
16 0732 CH12 4/24/01 11:24 AM Page 359
Figure 12.7
Sony Classical, a site that
Flash built. What it loses
in accessibility, it gains in
form and function.
Classical music greets
the visitor as the site
loads; as one piece of
music replaces another,
the company’s musical
offerings shift in the main
window as though one
were poring through
record bins (http://
www.sonyclassical.com/).
You can do some of this in
HTML and JavaScript but
not as smoothly or reliably.
For graphic designers, plug-ins are nothing new. If you want to create
strange blurs in Photoshop, you can buy and install Kai Krause’s KPT Filters
plug-ins. If you wish to work with preset masks in Photoshop, you’d pur-
chase Extensis Photoframe. Photoshop and Quark have sparked entire
industries devoted to creating such plug-ins. Plug-ins for web browsers
function exactly like plug-ins for Photoshop and Quark, except that
browser plug-ins are free. (Why are they free? How can they be free? We’ll
get to that.)
16 0732 CH12 4/24/01 11:24 AM Page 360
360 HOW: Beyond Text/Pictures: Turn on, Tune in, Plug-in
The other difference between designer plug-ins and browser plug-ins is

that browser plug-ins are as essential to the end-user as they are to the
creator. Your client does not need KPT Filters to see the way you’ve blurred
his logo in Photoshop, but web users need the Flash plug-in to view your
Flash work, the Real plug-in to see and hear your Real-encoded video, and
so on.
A Hideous Breach of Reality

Some plug-ins are true plug-ins, invisibly doing their work inside the
browser. Others are more like free-standing players, though their manu-
facturers still refer to them as plug-ins, and most web designers call them
that as well. Of course they should be called “helper applications” if they
aren’t actually plug-ins, as they have been for years by persnickety people
who also pronounce “GIF” correctly.
Still other multimedia add-ons can work either way. Depending on how you
mark up your web page, Apple’s QuickTime plug-in can lurk in the shad-
ows, invisibly playing embedded video and audio files right in the browser
window (like a plug-in). It also can spawn an overly ornate steel-burnished
console player that lies atop the browser window like a misbegotten Cadil-
lac ornament (not like a plug-in).
To keep it simple and to annoy the overly precise among you, we will fol-
low the manufacturers’ lead (as well as convention) and refer to all of these
add-ons as “plug-ins,” whether they behave like true plug-ins or not.
The ubiquity of plug-ins

Plug-ins have been a fact of web life and web design since the mid-1990s.
Why plug-ins? How did it happen?
In the beginning, God created the Heavens and the Earth. A little later, the
Web consisted of hyperlinked text. In the fullness of time, it became pos-
sible to include badly rendered images on web pages. This began to make
the medium more attractive to creative and commercial enterprises, and
there was soon a demand for sound, video, and other multimedia enhance-
ments on the Web. There was also a demand for really good Chinese food
in the American Midwest.
16 0732 CH12 4/24/01 11:24 AM Page 361
How to answer the clamor for sound and video and other fancy stuff? The
engineers at Netscape were inventing what a web browser was supposed
to be as they went along. Eventually they hit upon the notion of plug-ins.
If it worked for Quark and Photoshop, it ought to work for them too.
At the time, the Web was hotter than Jennifer Lopez’s Academy Awards
dress, and Netscape’s browser was the thing that was making it hot. Here
was a marketing opportunity! Dozens of plug-ins soon flung themselves
into the market. When new browsers began muscling in on Netscape’s turf,
they followed Netscape’s lead and supported “Netscape plug-ins” simply to
compete.
And here we are, more than half a decade on and still plugging away. (There
is still no good Chinese food in the American Midwest, however.)
THE IMPOSSIBLE LIGHTNESS OF PLUG-INS

Any web designer has much to say about plug-ins, not all of it printable,
and we are no exception. Truth is, we could write a whole book about plug-
ins. Come to think of it, we could write a book about just one plug-in. In
fact, many people have. Hillman Curtis, for instance, wrote Flash Web
Design: The Art of Motion Graphics, New Riders: 2000 (http://www.newrid-
ers.com/books/title.cfm?isbn=0735708967).
Want in-depth help with Flash, and penetrating insights into its nature?
Try Hillman’s book or Joshua Davis’s upcoming Flash to the Core (New Rid-
ers: 2001). We have our hands full as it is. Meantime, let’s assess a few
well-known plug-ins.
Plug-ins Most Likely to Succeed

There are as many plug-ins as there are stars in the heavens. Plug-ins for
specialty uses, plug-ins for novelty uses, plug-ins that support the needs of
mathematical and scientific markup, plug-ins that let you print official U.S.
postage from your web browser. No, we’re not kidding. Pour yourself a
strong beverage and hit http://home.netscape.com/plugins/ to take in the
range of available plug-ins. Do you have to learn about all these plug-ins?
No.
16 0732 CH12 4/24/01 11:24 AM Page 362
362 HOW: Beyond Text/Pictures: The Impossible Lightness of Plug-ins
When designing content that requires plug-ins, the first question to ask is,
which plug-ins? Which are most widely available? Which are likely to be
sitting in your visitor’s browser plug-in folder, just waiting for you to give
them something to play?
Unfortunately, this question is easier to ask than answer. An assessment of

which plug-ins come preloaded in which browsers does little to clarify the
state of plug-in-hood. As mentioned in Chapter 2, both Microsoft Internet
Explorer and Netscape Communicator include the Apple QuickTime plug-
in in their distributions. That much is known.
Netscape also includes the RealPlayer and Shockwave/Flash. IE for Win-

dows, the most popular browser/platform combination, includes the Win-
dows Media Player but not RealPlayer. Most Windows distributions of
Explorer include Flash (but not Shockwave); Mac Explorer distributions
include neither. By the time you read this, all of this might have changed.
That’s marketing, kids.
Seeing that you’re no better off than you were before, we’ve gone ahead
and created a short list of plug-ins we think you’ll run into during your
long and splendid web design career. Herewith, please review our never-
impartial assessment of the major multimedia plug-ins.
RealPlayer (www.real.com)
This popular plug-in/player delivers streaming video and audio, along with
support for Flash and the SMIL standard. RealPlayer is the most popular
streaming video format because it uses the least bandwidth and works on
all computing platforms (though it’s sometimes flaky on Macs).
As explained in Chapter 2, streaming video is video that plays while down-

loading. Early plug-in technologies did not stream. The viewer had to wait
for an entire movie or sound file to download before she could see or hear
the file in her browser. Real was the first to offer streaming playback.
Dull technical note

Even when a player supports streaming, a small amount of data must be
downloaded and cached before streaming begins to ensure smooth play-
back. If the file were to begin playing immediately, playback could be inter-
rupted later on—for instance, when other network traffic momentarily
16 0732 CH12 4/24/01 11:24 AM Page 363
interfered with the stream. By preloading (downloading and caching) ini-

tial data, the player attempts to offer smooth, uninterrupted playback.
Apple QuickTime and Windows Media Player, discussed in the following
section, are also streaming formats.
Tool tips
The free Real Producer software available at Real’s site is of sufficient qual-
ity for converting existing digital video to the Real format. An inexpensive
“Pro” version provides more options, enables you to create MP3 files, and
will help you create HTML files and stub files. Stub files are miniature text
files that trigger the streaming of Real files over conventional http net-
works. You also can easily create stub files yourself:
If lopez_dress.rm is a Real-encoded file featuring Jennifer Lopez in her

famous Academy Awards dress, and if that file resides at http://www.
example.com/real/lopez_dress.rm, to force the file to stream over a con-
ventional web server, you need to create a stub file. Open a blank text doc-
ument (Write, SimpleText, BBEdit, Tex-Edit, any basic text editor will do).
In this blank document, type the address of the Real file:
http://www.example.com/real/lopez_dress.rm
Close the document and save it as lopez_dress.ram. That’s your stub file,
Bubba. Upload it to your web server—for the sake of simplicity, we’ll assume
you’ve uploaded it to the same directory as the HTML file that references
it—and then conjure it up with a link like this:
<a href=”lopez_dress.ram”>La Lopez!</a>
Can you refer to it with an image instead? Of course. It’s just a link.
<a href=”lopez_dress.ram”><img src=”images/lopez_50x50.gif” height=”50” width=”50”
alt=”Actress Jennifer Lopez” title=”Watch Jennifer Lopez at the Academy Awards.
RealPlayer required.” Border=”0”></a>
You can also <EMBED> the file, thereby triggering the Real plug-in instead
of the Real console:
<embed src=” lopez_dress.ram “ autostart=”true” volume=”100” width=”2” height=”2”

controls=”hidden” pluginspage=”http://www.real.com”>
What does all that extra code mean? We’ll explain later.
16 0732 CH12 4/24/01 11:24 AM Page 364
The point is to link to the stub file, not the actual Real-encoded file. The
stub file, being text, downloads almost instantly. If the user’s browser is
configured correctly, the stub file will launch the RealPlayer, which will
then begin preloading the actual video clip. Your visitors will soon see Jen-
nifer Lopez…and so will Ms. Lopez’s attorneys. Be certain you have per-
mission to publish the clip.
You can skip the need for stub files if your client or host purchases dedi-
cated Real servers—recommended if you plan to serve much video to many
visitors. For instance, on a site that constantly serves TV news feeds, cable
comedy clips, or streaming video trailers, investing in Real servers would
be a Real good idea.
Special indications
With the highest compression ratio (and consequently, the lowest quality),
Real is the fastest streaming format, making it the plug-in of choice for
news sites and others where quality is less important than a hardy consti-
tution and the ability to deliver like James Brown. With Real, you are
assured of supporting the largest number of users with the widest range of
connection speeds.
If video files are meant to viewed and then forgotten (like a TV experience),
Real is likely what you want. Conversely, if video files are meant to be
stored and treasured on the end-user’s hard drive, you would probably
choose QuickTime instead. This is, of course, merely our opinion.
QuickTime (www.apple.com/quicktime/)
A high-quality streaming video format, QuickTime also supports a wide
variety of streaming audio formats including the ever-popular MP3. It also
delivers QuickTime VR panoramas and animated sprites. QuickTime sup-
ports hyperlinks in video, and offers some support for Flash and SMIL. An
innovative multistreaming process serves appropriately sized material
according to the end-user’s connection speed, as described in Chapter 2.
Last time we checked, QuickTime was the second most popular streaming
format. It uses more bandwidth than RealPlayer but delivers smoother
video and audio. The QuickTime format is native to Apple Macintosh com-
puters but thoroughly supports Windows PCs.
16 0732 CH12 4/24/01 11:24 AM Page 365
Tool tips
Delivering video over the Web should be impossible. The technology works
by means of drastic compression methods. Even the best of these methods
include visible artifacts. To minimize these artifacts and assure a better
quality image, choose subject matter that compresses well.
“Talking heads” compress well. Swish pans and mad action sequences do
not. This is because the compression works by seeking pixels that barely
change from frame to frame, choosing one of these pixels as a “master”
pixel, and repeating it from frame to frame in hopes of fooling your eye.
The more jerky camera movement in your mise-en-scene, the more obvi-
ous the software sleight-of-hand is. This is true not only for QuickTime but
for all digital video.
QuickTime VR panoramas can be breathtaking, and various software prod-

ucts are available to help you stitch together individual photographs into
a full 3D panorama, including Apple QuickTime VR Authoring Studio
(http://www.apple.com/quicktime/qtvr/authoringstudio/). You can insert
sound effects in these panoramas and confine the effects to certain por-
tions of the panorama. See Figures 12.8, 12.9, and 12.10 for nonprofit and
noncommercial QuickTime VR panoramas at Heidsite and PBS.org.
Figure 12.8
Writer Jim Heid’s personal
QuickTime VR panoramas
convey the rustic beauty
of his community (http://
www.heidsite.com).
16 0732 CH12 4/24/01 11:24 AM Page 366
Figure 12.9
A virtual reality tour of
Khufu’s Pyramid lends
needed “gee-whiz” appeal
to an essentially educa-
tional enterprise (http://
www.pbs.org/).
Figure 12.10
A hauntingly frozen
moment in time (http://
www.heidsite.com/). Print
absolutely cannot do this,
nor do these screenshots
begin to convey the effect.
16 0732 CH12 4/24/01 11:24 AM Page 367
For instance, a 360-degree panorama of the downtown skyline might be

filled with canned traffic effects until the viewer rotates the image to your
company’s office tower. Suddenly the sound of laughter is heard. As the
viewer rotates away from your office tower, the laughter is drowned by the
traffic noise.
Or a tranquil beach panorama could reveal the Jaws theme as the camera
reaches the sea.
Or a 360-degree panorama of the client’s company softball team could

reveal the hidden thoughts of each individual as the camera’s gaze passes
over him or her.
You get the idea. QuickTime VR can be very cool.
Special indications
If you want high quality, you probably want QuickTime. If you want VR
Panoramas, you need QuickTime. Other panoramic plug-in formats are
available (some of them quite good—see iPix, mentioned previously), but
none are nearly as widely distributed as QuickTime. Later we’ll talk about
how some plug-ins managed to crawl to the top of the heap while other
good ones languished. It isn’t really all that fascinating a story, but gossip
is as good a reading motivation as any, and her attorneys have informed us
that we can’t keep referring to Jennifer Lopez’s dress.
Windows Media Player (WMP)

(http://www.microsoft.com/windows/windowsmedia/)
WMP delivers streaming video and audio in the Windows Media File
(WMF) format and is included in all distributions of the various Windows
operating systems, making it a popular plug-in indeed. Though WMP is
viewed as a plug-in that primarily supports Windows users, a version is
available for Macintosh folk, and it actually works well. The player supports
real-time capture and broadcast of audio and video files (making it com-
petitive with RealPlayer) and also handles MP3 audio as smoothly as its
competitors.
16 0732 CH12 4/24/01 11:24 AM Page 368
According to Microsoft, Windows Media Player supports “near-DVD-qual-

ity” video and “near-CD-quality” sound. A free Windows Media Encoder
makes it as easy to prepare video materials for distribution as streaming
WMF files. As of this writing, the free encoder runs only in Windows 98 or
higher.
The WMP URL we’ve listed was accurate as this book went to press, but
contents might have settled during shipment. (Microsoft constantly
changes URLs at its site.)
Special indications
For some reason, WMF has apparently become the format of choice for the
streaming distribution of “adult” content, or so our informants tell us. We
just thought it was kind of interesting, that’s all. You try writing a chapter
about plug-ins, and see if your mind doesn’t start wandering where it
shouldn’t. We blame Jennifer Lopez’s dress.
Beatnik (http://www.beatnik.com/)
Musician Thomas Dolby’s Beatnik, though less widespread than the biggest
hitters, is an intriguing plug-in that comes bundled with most Netscape
distributions. Beatnik enhances MIDI (Musical Instrument Digital Interface)
playback, easily surpassing the quality of most PC add-on sound cards and
the Mac’s built-in MIDI voices. It also offers strangely wonderful features,
such as the ability to mix jam sessions on sites authored according to Beat-
nik’s specifications.
Such stuff is unlikely to be part of a site for your local church, synagogue,
or small business but might well add luster to the site for a recording artist.
Beatnik was used to enhance the PBS “Jazz” site (http://www.pbs.org/jazz/)
that accompanied Ken Burns’s historic jazz documentary series of 2001.
Beatnik works in Netscape (all platforms) and IE (Windows only). It unfor-

tunately does not work in IE5/Mac because its JavaScript functionality
relies on proprietary Netscape Application Programming Interfaces (APIs).
Beatnik is a cool plug-in, and IE5/Mac is an extremely polished, standards-
compliant web browser. It seems a pity that the two cannot work together,
but this raises the whole trouble with plug-ins, which we cover later in this
chapter (see “The Trouble with Plug-ins”).
16 0732 CH12 4/24/01 11:24 AM Page 369
Shockwave/Flash (www.macromedia.com,
www.macromedia.com/software/flash/)
Two plug-ins now bundled as one, Shockwave and Flash, are the biggest,
most-accepted, and possibly the most dynamic plug-ins on the market.
They are certainly the most controversial. Entire sites have been created in
these formats. Entire sermons have been written denouncing them.
For many, Flash is a religion; for many others, it is the first sign of the anti-
Christ. Flash artist Peter Balogh sums up the controversy in his witty essay,
“Sympathy for the Plug-in” at http://www.alistapart.com/stories/
sympathy/.
Figure 12.11
The “virtual piano” in the
Jazz Lounge at PBS.org,
created in Macromedia
Director, rendered unto
the Web in streaming
Shockwave format. Print
cannot do this (http://
www.pbs.org/jazz/).
Though they share a similar blow-you-away quality and though the man-
ufacturer now serves them like two peas in a pod, Shockwave and Flash are
quite different.
Shockwave
Shockwave is a sophisticated, proprietary format that can do anything a
CD-ROM can do. Full-fledged gaming environments; animated, hot-linked
city maps; endless labyrinths deep within simulated subterranean worlds:
If you can dream it up, it can be rendered in Shockwave. Essentially, Shock-
wave files are like self-contained software programs that launch in the
user’s browser.
16 0732 CH12 4/24/01 11:24 AM Page 370
As might be obvious, the more complex and multileveled the Shockwave

file, the larger it must be to do its job. Thus there is a trade-off between
sophisticated presentation and amount of bandwidth required.
There is also the risk that Shockwave programs will exceed the user’s com-
puting capacity. Linux users, who take justifiable pride in cranking tremen-
dous computing juice out of old, cheap PCs, frequently hit a wall when
Shockwave comes to town. Even Mac and PC users sometimes find Shock-
wave too rich for their blood. None of this is the fault of the operating sys-
tems in question. We’re back to the problem we discussed with Java. When
it works, it’s magnificent; when it doesn’t, it ain’t.
Shockwave files are created in Macromedia Director, a multimedia pro-

duction and programming package requiring tremendous expertise. No
web designer is expected to know how to program in Director, though some
specialize in it.
“Hip” web agencies generally have a Shockwave master or two in their

design departments—so do many “unhip” agencies. (We’re not sure what
“hip” and “unhip” actually mean in the context of web design and devel-
opment, but some web agencies seem to care a lot about it. In this, they
take their cue from ad agencies.)
Shockwave development is an art unto itself. It coincides with web design

but is not the same as web design. Shockwave has nothing to do with the
structured, semantic Web of meaning and information—but then neither
does a GIF image file.
Shockwave has largely escaped the fire and brimstone preached against its
younger cousin, Flash, because Shockwave files are fiendishly difficult to
create; therefore, gigantic Shockwave “intros” are not epidemic through-
out the Web. Hence the usability experts rarely scream about Shockwave.
But, oh brother, do they roar about Flash.
Flash
Flash is Shockwave’s lighter, less-bandwidth-intensive, easier-to-program
(but somewhat less powerful) cousin. Flash delivers animated graphics and
sound, and it is completely interactive. At this point in the history of the
Web, it is easier to create rich interactive presentations in Flash than by
trying to use the open standards of the Web (HTML/XHTML, CSS,
JavaScript, and the DOM). It’s more reliable, too, sadly enough.
16 0732 CH12 4/24/01 11:24 AM Page 371
Flash files stream and can be highly compressed because Flash is built on
vector graphics. (As you know, vector graphics, like PostScript, are mathe-
matical in nature rather than pixel-based.) Thus it is somewhat ironic that
Flash has become known as a bloated format. This has to do with poor Flash
authoring, not with Flash. Flash presentations, even incredibly sophisti-
cated ones, can be very low in bandwidth—and generally ought to be.
Flash files are created in the Macromedia Flash authoring program. They
also can be authored with Adobe’s LiveMotion software, which premiered
in early 2000.
Sixty second software review: LiveMotion is easier for beginners to learn,

particularly if they are familiar with the Adobe interface found in Photo-
shop, Illlustrator, and so on. It does one or two things Flash can’t do. Flash,
with a more baffling interface and a steeper learning curve, is initially
harder to learn. But its programming depth, through Action Script and
standard JavaScript, far exceeds that of LiveMotion (at least as of this writ-
ing).
Cult take: Flash designers take as much pride in mastering the tool’s
absurdly poor interface as they do in exploring its programming depth and
complexity. Just as web designers who code by hand take pains to chuckle
mockingly at Dreamweaver and GoLive users—and hold FrontPage users
utterly beneath contempt—so the hardcore Flash jockeys shake their heads
in bewilderment at the very notion that anyone would even think of using
LiveMotion.
Remember, this is not necessarily our opinion; it is mere observation of a

cultural milieu. If it sounds like Flash is a cult, it is. If it sounds like hard-
core web design is also a cult, it is. If it sounds like you should reconsider
getting involved in web design, you’ve read too far and know too much to
escape unscathed.
Our opinion: Flash and LiveMotion are both fine tools, each of which caters
to a different niche in the market. (There goes our cult status.)
Whichever authoring tool you use, you can do all the design work in the
program itself or create your vector graphics in Adobe Illustrator or Macro-
media Freehand and then export them to Flash (or LiveMotion). You also
16 0732 CH12 4/24/01 11:24 AM Page 372
might find yourself working with audio and video editing programs and 3D
design programs. If you think we’re going to cram this book with quick ‘n
easy tutorials on those sophisticated software programs, you are very high
and should lie down or consult a physician.
Tips on authoring Flash

Buy the program. Read the manual. What do you want from us?
Look at great Flash sites. Look at poor Flash sites. Emulate the good; learn
from the bad.
Good Flash sites are dynamic, attractive, navigable, intuitive, communica-

tive, and respectful of the visitor’s bandwidth and time—just like all good
websites.
Bad Flash sites are unresponsive, static, hard to navigate and understand,
communicate poorly (if at all), and waste the visitor’s bandwidth and time—
just like bad websites.
As with all such proclamations, there are always exceptions that succeed
in spite of being cryptic, initially confusing, or bandwidth-intensive. If you
are a genius with a deserved cult following, feel free to ignore the previ-
ous two paragraphs. If you’re not, respect your audience.
Choose vector graphics over raster graphics to conserve bandwidth. When

you must use raster graphics, use images that have been optimized to
death, rather than lovely images that suck bandwidth. A static four-color
GIF might not cut it on a traditional, static web page, but once that low-
grade image is set in motion, viewers will respond to the motion without
scrutinizing the quality of the image.
Juxt Interactive (http://www.juxtinteractive.com/) has built an entire prac-

tice by bringing brilliant design to Flash without the high bandwidth bag-
gage. Its prototype, the SHORN project (http://www.shorn.com/),
makes extensive use of four-color GIFs, and no one has ever noticed or
complained.
16 0732 CH12 4/24/01 11:24 AM Page 373
What Flash is great for

Flash excels as an environment for the creation of rich works of art
such as Monocrafts (see Figures 12.12, 12.13, and 12.14), Volume
One (www.volumeone.com/), Once Upon A Forest (www.once-upon-a-
forest.com), and many others you’ll meet in Chapter 13.
Figure 12.12
Those still perplexed by
the popularity of
Macromedia Flash need
look no further than
Monocrafts, Yugo
Nakamura’s multitiered
masterpiece.
Figure 12.13
Nakamura studied civil
engineering, architecture,
and landscape design
before focusing his crisply
uncanny intelligence on
issues of web art and
interface. These screen-
shots are from a previous
version of the site, which
still stands as a remark-
able achievement by any-
one’s measure but Jakob
Nielsen’s.
16 0732 CH12 4/24/01 11:25 AM Page 374
Figure 12.14
Reducing Monocrafts to a
series of printed screen-
shots is like trying to
explain a symphony by
playing a single note
(http://www.yugop.com/
ver2/).
What Flash is not so great for

Flash is not so great for structured data, semantic markup, accessibility,
searchability, indexability, and bookmarking. In short, nearly everything we
associate with the Web.
Why Flash gets a bad rap

Refer to the previous answer.
Why else Flash gets a bad rap

Refer to the previous answer.
Still more reasons why Flash gets a bad rap

There are several things, really. For one, too many copycat designers use
Flash in unimaginative, “me-too” ways. Thus, every other corporate site
seems to launch with a giant spinning logo rendered in Flash.
These miserable things are called “intros” (as in introduction) to the site.
They are the spiritual descendants of David Siegel’s (www.dsiegel.com/)
“entrance tunnels,” meaning that while some of them can be beautiful
establishers of mood, tone, and identity, too many feel simply gratuitous.
16 0732 CH12 4/24/01 11:25 AM Page 375
Here is a true horror story. The author of this book was supposed to speak
at a web conference. He visited the conference site in hopes that it would
tell him where the event was being held. He knew it was in a hotel, but
which hotel, and where? Instead of providing that information, the con-
ference site linked to the third-party site of the hotel itself (strike one). Not
the most usable idea, but all right.
The hapless author clicked the hotel site link and discovered that he was
trapped inside the conference site’s frameset, an HTML error so basic he’d
forgotten such things could even happen. Strike two.
Inside that tiny frameset, the hotel site presented, not its name or address,
but a Flash intro (strike three), nor was this an optimized, bandwidth-
friendly intro built largely with vector graphics. No, what it actually was,
was something more akin to a QuickTime movie: a motion picture showing
the beauty of the hotel. Strike four. In spite of his fast connection, the
author had to wait several times while the overweight Flash file choked on
its own girth before streaming resumed. Strike five. There was no way to
skip this intro; it simply had to be endured. Strike six.
When the low-resolution but high-bandwidth graphic nightmare finally

ended and the expiring Flash file triggered a standard HTML page, that
page did not list the hotel’s address. Strike seven. The author had to navi-
gate three layers deep into the site before he could find that simple, essen-
tial bit of information. In baseball you get three strikes. On the Web, you
might not even get one. We needed to locate that hotel’s address because
we had no choice but to show up there. Travelers planning a trip have many
hotels to choose from. Word up.
And that’s another reason that Flash gets a bad rap—because people who
don’t know what they’re doing often use it poorly. There are bad painters,
but no one criticizes paint. On the other hand, bad paintings aren’t shoved
in your face when you’re trying to find information online.
A “Flash usability” site at www.flazoom.com can help you avoid designing

useless or less-than-usable Flash movies; beyond that, the issues we dis-
cuss throughout this book are equally applicable to the world of Flash. Base
your work on the needs of your audience; create intuitive structures that
invite your specific audience to enter, explore, and linger; craft a memo-
rable identity without wasting bandwidth…you know the drill by now.
16 0732 CH12 4/24/01 11:25 AM Page 376
376 HOW: Beyond Text/Pictures: Who Makes the Salad?
Clients might salivate over the prospect of giant animated logos, but
designers and web users are tired of this unimaginative use of Flash, and
the plug-in is best reserved for truly creative and artistic purposes. You’ll
find sites exemplifying that kind of creativity in Chapter 13.
Another reason Flash gets a bad rap

Failure to provide alternatives, thus leaving some users in the dust, is a
widespread problem. Macromedia has begun an initiative to make Flash
sites more accessible, and this is commendable. But streaming audio and
visual media, accessed by mouse movements, will always remain inacces-
sible for some (refer to the preceding discussion of VRML).
Unless Flash is able to overcome tremendous barriers to accessibility

(http://www.alistapart.com/stories/unclear/) inherent in its very nature, it
is vital to provide some kind of basic HTML alternative for those who might
be unable to see or hear or move their limbs.
Embedding Flash files in web pages

Flash generates all the HTML necessary to embed a Flash movie in a web
page. Isn’t that nice? You might, of course, want to go in and further mas-
sage the markup.
Are we finished with Flash?

We will never be finished with Flash, but we’ve pretty much finished dis-
cussing it in this chapter.
WHO MAKES THE SALAD? WEB DESIGNERS

AND PLUG-INS
As a web designer, you will rarely be asked to develop content for plug-ins
(with the possible exception of Flash). QuickTime and Real video files, for
example, are usually created by web producers or design technicians—not
by web designers. Sometimes the people who actually shoot and edit the
film will generate digitized versions as part of the post-production process.
16 0732 CH12 4/24/01 11:25 AM Page 377
If you do find yourself pressed into service in these areas, the work is not
hard. In agencies, you’ll typically find a workstation with a video machine
at one end and a Mac or PC at the other. Software like Adobe Premiere or
Apple iMovie is used to digitize the film. Third-party compression software
such as Terran Cleaner 5 (http://www.terran.com/) optimizes video for
streaming web delivery. None of this is difficult, but rarely will a web
designer be asked to do it. Your time is too valuable elsewhere.
Many web designers include Flash design in their skills repertoire; many
others do not. Developing exceptional Flash content is a specialization all
its own. Most web agencies keep a few designers on staff who excel at
Flash development, allowing the bulk of the design crew to focus on inter-
face and other design issues.
So if web designers generally do not create plug-in content, what do web

designers have to do with plug-ins? They make them work on web pages—
that’s what.
Making It Work: Providing Options

Web designers use HTML to embed a plug-in file (or object) on a web page.
Following is markup from an IPIX panorama page at the Travel Channel
(http://travel.discovery.com/tools/gallerypages/ipix/noam/wmiss/
zabptcaj.html):
<p>

width=”280”>
</applet>


<noembed>
<b>To view IPIX images you need to Upgrade your Browser. We recommend Version 3
or above of either Netscape or MSIE.</b><br>
</noembed>

</p>
16 0732 CH12 4/24/01 11:25 AM Page 378
Notice the use of  to help the web designer keep
track of what she is doing and why.
Notice that room has been made for the space used by the plug-in image
file (280 x 210 pixels). If the layout for this web page had been initially cre-
ated in Photoshop, the web designer would have left a 280 x 210 space in
the layout itself and then replaced it with HTML during the web-building
phase. In all probability, this page was never individually designed in Pho-
toshop but is simply one of many that share the same template.
Notice that a Java applet has been used to embed the file, as described in
this chapter’s section on Java. Visitors lacking the iPIX plug-in will be
treated to a Java simulation, rather than simply encounter an error mes-
sage about missing plug-ins.
But what happens to the visitor whose browser does not support either
plug-ins or Java? That is taken care of by the <noembed> tags. Let’s look
just at that section of the markup:

<noembed>
To view IPIX images you need to Upgrade your Browser. We recommend Version 3 or
above of either Netscape or MSIE.<br>
</noembed>

</p>
The <noembed> tag basically says, “Listen up, old, dumb browser. What fol-
lows is for you.” In such a browser, the text beginning To view IPIXimages…
will be revealed. In other browsers (those that support Java or contain the
iPIX plug-in), the text message will be hidden.
The designers also could have put this text and markup inside the <applet>
element, like so:

width=”280”>
<p>If you can read this, your browser does not support Java. Have a nice day.</p>
</applet>
16 0732 CH12 4/24/01 11:25 AM Page 379
As to whether anyone ever upgraded their browser in direct response to a

website’s message, well, that is something else again.
The “Automagic Redirect”

Whether bundled with the browser or not, all plug-ins are readily available
online. In some cases, when a web user hits a page that requires a plug-in
not found on her system, Java is substituted for the missing plug-in (as in
the preceding example).
In most cases, though, Java is not pressed into service. After all, plug-in
manufacturers want their plug-ins to be downloaded, not synthesized by
a substitute technology. Typically, when a web user lacks a plug-in, she is
“automagically” directed to the appropriate plug-in page so she can down-
load it. In most cases, the magic is merely a matter of adding a <plug-
inspage> attribute to the HTML <OBJECT> or <EMBED> tag:
<embed src=”http://build.kubrick.org/sounds/the.shinning/midnight.the.stars.and.you.mp3”
autostart=”true” volume=”100” width=”2” height=”2” controls=”hidden”
pluginspage=”http://www.quicktime.apple.com”>
In this snippet lifted from our unfinished experiment at www.kubrick.org,

the <pluginspage> attribute to the <EMBED> tag serves two functions:
1. The material in question is an MP3 audio file (midnight.the.stars.

and.you.mp3), capable of being played by a number of additional
plug-ins and players, including RealPlayer, Flash, and a variety of free
and commercial applications. Specifying QuickTime in the <plug-
inspage> attribute tells the browser which plug-in to select: namely,
Apple’s QuickTime plug-in. More about why that’s important a little
later in this section.
2. It also provides a web address where the latest version of the Quick-
Time plug-in can be downloaded. If the visitor did not have Quick-
Time installed on her system, the browser would display a dialog box
indicating that the plug-in was required and asking if she wished to
download it. Clicking Yes or Okay would load the appropriate Apple
plug-in page. The way the <EMBED> tag works, the browser does
most of the work of supporting the web user (and the web designer).
16 0732 CH12 4/24/01 11:25 AM Page 380
Embed 'n Breakfast

<embed
src=”http://build.kubrick.org/sounds/the.shinning/midnight.the.stars.and.you.mp3”
autostart=”true” volume=”100” width=”2” height=”2” controls=”hidden”
pluginspage=”http://www.quicktime.apple.com”>
Heck, while we’ve got this markup in front of us, let’s just go ahead and
explain what the rest of it means:
■ <EMBED>. This tells the (2.0 or higher) browser to anticipate content

that must be handled by a plug-in.
■ <AUTOSTART>. This tells the browser to begin playing the file instantly.
(The default value is on.)
■ <VOLUME>. This sets the loudness. (The default value is 100, or full
volume.)
■ <CONTROLS>. This specifies the presence or absence of on-screen con-

trollers, similar to those on a video or audio cassette console. If controls
are visible, they can be seen and used by the visitor. If hidden, they do
not appear, and consequently they take up no space on the screen. (The
default value is visible.) When would you use hidden? You’d use it, as we
have here, when you simply want the file to play without prompting the
visitor to do anything. Naturally, in such a case, you’ll turn Autostart on.
Otherwise, you’re forcing the user to download a file they have no means
of playing.
■ <WIDTH>, <HEIGHT>. These attributes specify the size of an in-page

controller, if any. Interesting paradox: If controls are hidden, why specify
sizes at all? It’s because Netscape 2, 3, and 4 might crash if some size
attribute were not included. It also could crash if the size were smaller
than 2. As to the number itself, “2” means “2 pixels.” Though width and
height are specified, they do not appear because the “hidden” value of
the <CONTROLS> attribute makes them invisible. If you’re working with
hidden player controls, the default width and height attributes should be
2 to avoid crashing old versions of Netscape Navigator.
16 0732 CH12 4/24/01 11:25 AM Page 381
Before we close this fascinating portion of our narrative, we must add one
more reason to specify the player via the <pluginspage> attribute: If you
don’t, the browser will choose one for you, often with hideous results.
Read on.
The iron-plated sound console from Hell

Right up through its 4.0 browser, Netscape used to respond to WAV, AIFF,
AU, and other traditional sound file formats by sprouting an ugly little con-
sole. But the console did not simply leap up and start playing. Oh, no. Nor
was the console actually part of Netscape’s browser, even though it was
the default player. For reasons we can only guess at, Netscape chose Java
as the foundation for the console.
When you encountered a site that contained a sound, the page would stop
loading, and the browser would seem to freeze. In the status bar, the
dreaded words “Starting Java…” would appear. After a Vietnam-like eter-
nity, the ugly console would at last pop up and blast the stupid sound.
Now, suppose you did not feel like waiting for this mockery of a sham to
run its course. Suppose you attempted to close the browser window or nav-
igate to a previously visited site via the Back button. What would happen
then? The browser would crash, of course.
If most people did not detest embedded sound files to begin with, this
tragicomic exercise in non-user-centric design certainly encouraged them
to think of embedded sounds as one of Satan’s more diabolical efforts.
THE TROUBLE WITH PLUG-INS

While providing the visitor with linkage to the appropriate plug-ins page
is certainly a friendlier gesture than simply abandoning her to chance, most
professionals try to go one step further. They try to hide all the technolog-
ical complexity from their users. Even something as simple as navigating
to a plug-ins page can confuse and frustrate some users.
16 0732 CH12 4/24/01 11:25 AM Page 382
382 HOW: Beyond Text/Pictures: The Trouble with Plug-ins
To work around this, most developers step in at this point and write a plug-
in detection script. The theory is simple: If the user has the plug-in, the
embedded content plays. If the user lacks the plug-in, some alternative is
provided (perhaps something as simple as text). The user is never made to
feel inadequate, never made aware that she might be missing something.
It’s a beautiful plan, but as we mentioned in the JavaScript chapter, it has

often broken down because plug-in detection is not universally supported.
Netscape, having created JavaScript, has always used it in the browser to

detect the presence or absence of plug-ins. Let’s take the Flash plug-in for
argument’s sake. If the plug-in is not detected, the visitor might be taken
to a page that explains that the site uses Flash and offers her the oppor-
tunity to download the plug-in from Macromedia.com, as previously
described in “The ‘Automagic Redirect.'”
Because JavaScript was not originally a standard technology, Microsoft’s

Internet Explorer had to rely on another technique. Prior to IE5, Microsoft
used IE-only ActiveX technology to handle plug-in detection.
Before writing plug-in detection scripts, developers had to write browser

detection scripts. If the browser was Netscape’s, the JavaScript plug-in
detection script ran. If the browser was IE, ActiveX plug-in detection was
triggered (and if the plug-in was missing, ActiveX would supply it).
None of this worked on the Macintosh version of Explorer, whose users

generally ended up in a hellish loop of nonfunctioning technology and self-
contradictory error messages. This cruel stupidity should not be blamed on
the Macintosh Operating System, nor on developers who toiled long and
hard to work around browser deficiencies.
IE now supports JavaScript on both the Windows and Macintosh platforms.

As users upgrade to new versions of these browsers, these incompatibility
problems should become a distant memory.
Yet software developers still sometimes confuse Netscape’s proprietary

JavaScript APIs with standard JavaScript. That’s why two plug-ins men-
tioned earlier in this chapter (Adobe’s SVG plug-in and Thomas Dolby’s
Beatnik plug-in) don’t work properly with IE5/Mac.
16 0732 CH12 4/24/01 11:25 AM Page 383
And web designers who don’t keep track of the ever-changing browser
compatibility scene still make silly mistakes, particularly where IE5/Mac is
concerned. For instance, even though IE5/Mac handles plug-in detection
flawlessly, many Flash sites, when they detect the presence of IE on a Mac,
refuse to let the user proceed until she has switched to Netscape’s browser.
This makes no sense, but it happens all the time.
We fear we are beginning to lose some of you in the back row. Snap out of
it. We’re almost done, honest.
If Plug-ins Run Free

Earlier we promised to answer a simple question: Why don’t companies
that make plug-ins charge web users to download them? After all, Exten-
sis makes a bundle from its fine Quark and Photoshop plug-ins. Are the
makers of the most popular plug-ins (Macromedia, Apple, Real, and
Microsoft) simply beautiful altruists who want to teach the world to sing
and don’t desire a penny for their efforts?
That was, of course, a rhetorical question.
Companies distribute their plug-ins at no cost because the value of these

products is commensurate with their distribution. Put simply, a plug-in that
is on 100 million desktops is vastly more valuable than one that is on a mil-
lion. How do you encourage a person to try something? Let them have it
for free.
Indeed, as we’ll see in a moment, companies not only gave stuff away free,
they paid other companies to promote their free stuff. Never have so many
spent so much to earn so little. (Excluding the browser wars themselves, of
course. Those cost even more and made even less.)
Okay, so as a result of giving all this good stuff away for free, Macromedia,
Apple, Real, and Microsoft have achieved what they sought: nearly every-
body uses their plug-ins. So how do these companies recoup their invest-
ment and hopefully even squeeze out a profit?
They do it by creating and selling authoring tools. Web designers buy

Macromedia Flash. Web producers buy Real Producer and professional
QuickTime authoring suites.
16 0732 CH12 4/24/01 11:25 AM Page 384
384 HOW: Beyond Text/Pictures: Parting Sermon
Though Apple sells professional QuickTime suites, it also gives away some
extremely capable video authoring tools with every new Mac. What is the
sense in that? The sense in that is that these “free” products come with a
Macintosh. If you want the free product, you buy the Macintosh computer.
Similarly, Microsoft gives away WMF authoring tools to encourage you to
buy Windows products. Some web businesses might have trouble coming
up with revenue models, but software and computer companies generally
don’t.
While striving to reach ubiquity, plug-in makers have frequently partnered

with content producers. For instance, at different times, downloadable
trailers at a well-known movie company’s empire of websites have been
available exclusively in Apple’s QuickTime format, and at others, exclusively
in Windows Media Player format. The plug-in maker compensated the
movie studio for favoring its product over competitive plug-ins.
Today the cash flows in the opposite direction. A movie studio might pay
the purveyor of a popular plug-in to feature its studio instead of a com-
petitive film conglomerate on the plug-in vendor’s “Hot Downloads” page.
Ubiquity makes for destinations, and destinations, if popular enough, can
generate income. To up the income, the plug-in page sprouts ad banners—
from free plug-ins to cold cash in twelve easy lessons.
As we feared, none of that was as interesting as Jennifer Lopez’s dress.
PARTING SERMON
In Chapter 2, we discussed the way most popular plug-ins stream their pre-
sentations to compensate for slow user dialup modem speeds. We also
reminded you just how slow those dialup speeds really are. Please reread
Chapter 2 before authoring high-bandwidth multimedia content or blithely
adding it to a site for which it might be inappropriate.
We web designers, most of us, anyway, live in a spoiled world of hyper-fast

Internet access, powerful desktop processors, and wide-screen monitors.
Most of the world does not enjoy such niceties, or anything half so nice.
16 0732 CH12 4/24/01 11:25 AM Page 385
In fact, as the Web grows in popularity, the median average access speed
declines drastically because there are more and more home users for every
luxuriously appointed web professional. Though the field is expanding ten-
fold, the web-using population is growing at many times that pace. Even
if the profession were to stop growing, the number of web users would
continue to rise.
The day of universal high-speed access and fat bandwidth is not at hand.
It’s not even close.
While the prophets of high bandwidth high-five each other, millions in

China and Africa and Alabama begin using the Web via a 14.4 modem that
is shared by two or three families or 50 kids in a schoolroom. In libraries in
America and around the world, those who cannot afford Internet access
line up for hours to use public systems. Some of those systems are fast.
Some are not. Few can afford to be tied up for hours just so some logo can
spin.
Even those on the fortunate side of the digital divide rarely enjoy the
fastest speeds or the most reliable connections. When the Daily Show’s Jon
Stewart jokes about the AOL busy signal, the entire audience laughs.
They’ve been there. Most of them still are there. They are not only Jon
Stewart’s audience, they are every web designer’s audience. And they’re the
ones in the good seats.
So treat rich media like you’d treat Jim Beam: responsibly.
We end this chapter on a somber note, but the book on a happy one. Kindly
proceed to Chapter 13.
16 0732 CH12 4/24/01 11:25 AM Page 386
17 0732 CH13 4/24/01 11:25 AM Page 387
chapter 13
Never Can Say Goodbye
YOUR DIALOG WITH THE WEB has now begun. And though this book, like young
love, must end, our conversation will continue. You will find us, and we will
find you on the pages of the World Wide Web.
No book (indeed, no five-year program, if one existed) could teach you

everything you need to know to design smart, attractive, user-focused
websites. You will learn as you work—from teammates, partners, and even
your clients.
You also will learn a great deal from the people who visit your sites. You’ll
be surprised at how many write—and not merely to complain when your
single-spaced, 10px type sends them scurrying to the optometrist.
But some of the best places to learn are on the Web itself, hence this chap-
ter. In it we share our favorite online resources and explain the importance
of continuing your education as the Web and your career experience
growth and change.
SEPARATION ANXIETY
Throughout this book, we’ve shown methods used to design today’s Web
and shared theories about how people interact with the medium. You need
to know these things to begin working now.
17 0732 CH13 4/24/01 11:25 AM Page 388
388 HOW: Never Can Say Goodbye: Separation Anxiety
But as we’ve also pointed out, the Web is changing; indeed, like the sea, or
like some other Zen metaphor we can’t quite put together here, the Web’s
very nature is one of constant change. Currently the Web is changing in an
intriguing way—one that will move it closer to its founders’ original vision
of an open medium, accessible by all people and available to all sorts of
Internet-enabled devices.
What will empower that happy change? It will come with the separation
of style from content. What does that mean? It means you’ll stop welding
your texts and functions and images together through overextended HTML.
Instead, you’ll keep your visual design in one place (a Cascading Style
Sheet) and your content in another (a series of HTML or XHTML documents;
a database of XML-formatted text). The twain will meet on the web page,
but their behind-the-scenes separation will considerably enhance your
working conditions and your audience’s experience.
Instead of painstakingly slicing apart images in Photoshop as described in

this book or spending hours hand-tweaking hundreds of individual HTML
documents, you’ll have time to spend on more interesting pursuits such as
design itself—which is, after all, what you do.
This change in the nature of web design as a practice will come when all
web users employ browsers that fully support the standards that empower
us to separate style from content: HTML/XHTML, CSS, XML, JavaScript/
ECMAScript, and the DOM.
Not only do browsers have to change (and they are changing), web design-
ers must also change—a proposition that requires the willingness to con-
tinue learning and to risk discarding methods we’ve spent years perfecting.
In February 2001, A List Apart reinvented itself with a standards-compliant

design that separates style from content (http://www.alistapart.com/
stories/99/). As you might expect, the site (www.alistapart.com) is a good
resource for information on that subject.
The reinvention of ALA coincided with The Web Standards Project’s Browser
Upgrade campaign (http://www.webstandards.org/upgrade/), which urges
web designers to learn about and use the W3C recommendations we’ve
17 0732 CH13 4/24/01 11:25 AM Page 389
discussed in this book, even if the resulting sites look less than delicious in
older, nonstandards-compliant browsers. The Browser Upgrade campaign
also asks web designers and content creators to seek ways to encourage
user upgrades so that the Web can improve without leaving anyone behind.
The Browser Upgrade campaign and the ALA redesign were logical next
steps in the evolution of the Web. We launched them while writing this
book, which brings up the problem with books. Namely, while books have
the virtue of permanence, they cannot update themselves as websites can.
We encourage you to continue learning by visiting educational and inspir-
ing websites and reading and participating in web design mailing lists and
forums.
The remainder of this chapter will provide you with plenty to choose from.
Use these resources to amplify parts of this book and to learn more about
the emerging, standards-based Web. At the end of the annotated list below,
we’ll return to offer a final thought about the Web and you.
FROM TAG SOUP TO TALK SOUP: MAILING

LISTS AND ONLINE FORUMS
Learning by trial and error is part of any process and is certainly part of web
design. Learning from other members of your team is a deeply bonding
experience, but learning (and sharing your own knowledge) on a mailing
list is a pleasure no web designer should miss.
There are many, many mailing lists and online communities for web design-
ers and developers. Some focus on specific technologies; others are vast,
crowded, and general. Some function as job referral services while others
mainly promote the people who created the list. Some are chaotic, others
restrictive. With a little effort, you will find the ones that make you feel
most comfortable.
Following, in alphabetical order, are some of our favorites.
17 0732 CH13 4/24/01 11:25 AM Page 390
390 HOW: Never Can Say Goodbye: From Tag Soup to Talk Soup
A List Apart
http://www.alistapart.com/
Each week A List Apart publishes useful tutorials (“Meet the DOM,” “Fear
of Style Sheets”), challenging opinion pieces (“The Curse of Information
Design,” “Sympathy for the Plug-in”), or both. And each week, after read-
ing these articles, ALA readers respond on the site’s discussion forum. The
site is noncommercial, and you need not reveal your identity or other per-
sonal information to participate in the discussion forums.
Astounding Websites
http://www.astoundingweb.org/
Launched by Glenn Davis and maintained by Dave Bastian, this unique dis-
cussion community was created to honor the best writing, design, and pro-
gramming on the Web. Visit this small, friendly forum to discover inspiring
commercial and noncommercial sites or participate by reviewing sites you
admire. You can also submit your own sites for review in the Site Promo-
tion section.
The Babble List

http://www.babblelist.com/
Maintained by Christopher Schmitt (and resurrected by him in 2001 after

a brief hiatus), The Babble List is a well-run general web design mailing list,
covering issues of graphic design, information architecture, writing, usabil-
ity, project management, and related skills. Though the average Babble Lis-
ter is a professional with at least two years’ experience, the list is
beginner-friendly. If you find yourself stuck on a JavaScript or CSS prob-
lem or wondering why your site looks great in one browser but poor in
another, you can post your message to The Babble List and anticipate use-
ful feedback.
17 0732 CH13 4/24/01 11:25 AM Page 391
Dreamless
http://www.dreamless.org/
Dreamless is a deep and open community primarily populated by young

graphic designers and Flash artists. Though the site’s gray-on-gray, Arial-
only design gives it a somber appearance, it’s anything but dull. Dreamless
discussions range from the seriously spiritual to the deliberately silly. The
site has a fanatical following and encourages its members to get together
at parties in various cities. If you have trouble finding the site’s front door,
use View Source.
Evolt
http://www.evolt.org/
Evolt, a multi-faceted mailing list, online message board, and member-cre-

ated publication, provides useful dialog spaces for technically minded web
designers and developers worldwide. Accessibility and web standards are
hot topics here, and you can learn simply by reading other members’ posts.
Like all communities mentioned here, Evolt is self-policing; and like all suc-
cessful communities, it manages the task unobtrusively.
Metafilter
http://www.metafilter.com/
Matt Haughey’s noncommercial community site is not about web design

or web programming, but many web content creators will be found in its
forums. Billing itself as a “community weblog,” the occasionally raucous
discussion site can help you get a handle on aspects of the Web’s emerg-
ing culture. This in turn will remind you that the Web is not about HTML
tags or graphic design; like Soylent Green, the Web is people.
17 0732 CH13 4/24/01 11:25 AM Page 392
392 HOW: Never Can Say Goodbye: From Tag Soup to Talk Soup
Redcricket
http://www.redcricket.com/
Dan Beauchamp’s personal site includes a web design forum (“Commu-

nity”) that’s small, lively, and friendly. HTML questions? JavaScript woes?
Redcricket could be the ticket. By maintaining a fairly low profile, Red-
cricket’s forum generally avoids the flame wars and ego trips that some-
times plague other lists and communities. Spend time at the site before you
post. Redcricket is a tight community of friends; barging in and loudly
demanding attention won’t go over well.
Webdesign-l
http://webdesign-l.com/
Stewarded by Steven Champeon, Webdesign-l is a long-running, smartly

focused design and development list. Some of the brightest people in the
industry participate in this highly respected list. Champeon, a systems guru
who technical-edited Taking Your Talent to the Web and who co-founded
The Web Standards Project, runs a tight ship. As list administrator, he keeps
misinformation to a minimum and stops bad behavior before it starts.
Beginner questions might be well-received if submitted with restraint.
(“Hellllllp! My site is hosed!!!!!!” will probably not generate the kind of
feedback you want.) Read the list rules and get used to the general dis-
cussion tone before posting to the list.
When All Else Fails

http://www.r35.com/edu/
Consider a class. R35edu offers a curriculum of over 60 courses, covering

nearly every facet of web strategy, design, development, commerce, and
marketing—all via “a unique distance learning environment that puts you
in direct contact with creative innovators and designers from all over the
world.”
17 0732 CH13 4/24/01 11:25 AM Page 393
EYE AND BRAIN CANDY: EDUCATIONAL AND

INSPIRING SITES
Attempting to figure out web design exclusively from a book is like trying
to learn about music without listening to any. Fortunately, the Web is rich
in inspiring and educational sites. Following are a few of our favorites,
including a couple of our own (cough).
Design, Programming, Content

A List Apart (http://www.alistapart.com/), “for people who make websites.
From pixels to prose, coding to content.” See previous section for more
on this.
Apple Internet Developer (http://developer.apple.com/internet/), launched

in 2001, started small, but what it has is choice: brief and pungent tutori-
als on HTML, online typography, CSS, JavaScript, and the DOM.
Builder (http://www.builder.com/), “solutions for site builders,” provides

articles and tutorials on graphic design, multimedia, back-end develop-
ment, and even software (“Fireworks vs. ImageReady”). There is also a dis-
cussion board (Builder Buzz), and the site hosts a dandy annual web design
conference in New Orleans.
Each month, Digital Web (http://www.digital-web.com/), “the web

designer’s online magazine of choice,” brings you fresh interviews, tutori-
als, columns, and even classifieds (to help you get your next job). Edited
and published by Nick Finck, who also contributes to A List Apart.
Web Page Design for Designers (http://wpdfd.com/), published monthly by

Joe Gillespie, is “aimed at people…already involved with design and typog-
raphy for conventional print, [who] want to explore the possibilities of this
new electronic medium.” In other words, it speaks to the audience of this
very book! (We would have titled this book “Web Design for Designers” if
Joe hadn’t beaten us to the punch, darn him.) The site includes typefaces
optimized for the Web, columns on web design and typography, and a solid
listing of third-party resources.
17 0732 CH13 4/24/01 11:25 AM Page 394
394 HOW: Never Can Say Goodbye: Eye and Brain Candy
The Web Standards Project (http://www.webstandards.org/), co-founded

by Glenn Davis, George Olsen, and your humble author, maintains a
Resources section for your educational pleasure. Confused about CSS,
ECMAScript, and the rest of the alphabet soup? You’ll find links to relevant
articles here.
Web Techniques (http://www.webtechniques.com/) is a vast, professional

publication with an accompanying real-world magazine you can read in
the bathtub or carry in your attache case. It covers web technology and
business and can help you understand how wireless technology interfaces
with web design.
Web Review (http://www.webreview.com/) publishes some of the smartest

tutorials we’ve ever seen on XHTML, JavaScript, and other web technolo-
gies and has always been a great friend to web standards. Highly recom-
mended, particularly for those who wish to understand web technologies
instead of simply pushing buttons in WYSIWYG editors.
Think of Webmonkey (http://www.webmonkey.com/), originally directed by

Jeff “Art & Science of Web Design” Veen, as Builder.com with more atti-
tude. A deep resource dating back to the earliest days of the designed Web,
the site sports swell tutorials on HTML, JavaScript, and other technologies,
along with columns and articles on streaming media, emerging standards,
and the web business. Not updated as often as it used to be, but still a fine
smoke.
Webreference (http://www.webreference.com/), a subsidiary of

Internet.com (yes, there really is an Internet.com), is tailored more
to developers than designers but will repay your exploration. Edited by
Andy King, the vast site covers everything you could ever want to know on
the web technology front. Interviews and discussion forums enhance the
site’s value.
Webtype (http://www.webtype.org/), dedicated to better online typogra-

phy, keeps you posted on this vital and sadly under-reported topic. (Some-
times web designers seem more interested in scripting and gimmicks than
they are in ensuring that type is legible—let alone attractive and pleasur-
able to read.) Webtype gives you the lowdown on everything from
17 0732 CH13 4/24/01 11:25 AM Page 395
readability studies and CSS nuances to typographic explorations and

downloadable typefaces. Don’t miss the survey of fonts installed on PC and
Mac users’ computers. Founded by the mysterious “Gen,” with kibitzing
from Dave Bastian, Joe Clark, Julia Hayden, Webmistress Jo, and your hum-
ble author.
The World Wide Web Consortium (http://w3.org/), the mother of us all, is

the final authority on web standards. Use it to keep track of existing and
emerging technologies and to verify the way these technologies should
work, before running off half-cocked, screaming about aliens jamming the
radio transmitter embedded in your skull as part of an evil CIA experiment.
Note that W3C articles, while definitive, are among the least easy to read
and understand of any we’ve seen—and that includes VCR manuals writ-
ten in Japan. You’ll do better if you check W3C to see what you should learn
about; then read the friendly tutorials at Webmonkey, Builder, or A List
Apart.
The Big Kahunas

Let us now praise famous art directors:
Adobe (http://www.adobe.com/) not only makes great software for print

and web designers, they also run a fine, vast site full of tutorials, columns,
and articles on web, print, and motion design. Disclaimer: Your humble
author writes a column for this publication.
AIGA (http://www.aiga.org/), the American institute of Graphic Arts, has a

long and noble history as a membership organization for designers. But you
know that. The site helps you track seminars and conferences and offers a
national job bank and member discussion board along with thought-
provoking articles (“What is Graphic Design?”).
Communication Arts (http://www.commarts.com/) is among the world’s

most-respected voices for design. Its interactive section includes design
technology columns and a Website of the Week. And of course the Com-
munication Arts annuals honor some of the best design and advertising
communications in the world.
17 0732 CH13 4/24/01 11:25 AM Page 396
PDN-Pix (http://www.pdn-pix.com/pix/), the digital arm of Photo District

News, provides web design features (“Waiting to Load”), Q&A (“Ask Pix”),
reviews of noteworthy sites (“Pix’s e-Projects”), and a column by your hum-
ble author (“Second Site”). The print magazine will repay your interest;
much of this material gets republished on the site along with some web-
only content (“Grand Masters of Flash”).
Beauty and Inspiration

When grinding out menu bar buttons saps your inspiration, trust well-
designed, meaningful sites to restore it. Begin your voyage with sites that
deliver compelling, original content (and not in plain brown wrappers):
{fray} (http://www.fray.com/), the ultimate personal storytelling site, was

conceived, produced, and art-directed by designer/author Derek Powazek.
In addition to showcasing what an imaginative web designer can do with
words and pictures, the site functions as an on and offline community.
Highly recommended. (Derek Powazek is also the author of Design For Com-
munity, published by New Riders.)
Glassdog (http://www.glassdog.com/), Lance Arthur’s personal magnum

opus, is both sarcastic and smooth. As if the site’s clever writing and smart
scripting were not intimidating enough, Arthur manages to combine clean,
spare, easy-to-navigate design with the technical dexterity of a dazzling
showoff.
Harrumph! (http://www.harrumph.com/), Heather Champ’s charming and

witty online diary, sports one of the cleanest web layouts we know. Perhaps
this is because Heather has been designing websites since 1995, or perhaps
it’s because she’s got taste. All we know is, every site that uses words
should be this easy to read and engaging to look at. Few are.
Media.org (http://www.media.org/), “a collective of artists/architects…

fueled by a passion for the Internet medium,” was cofounded by Carl Mala-
mud and Webchick in 2000 to debunk web inanities, promote web intelli-
gence, and rescue digital works laid waste by careless businesses. Among
the sites they rescued:
17 0732 CH13 4/24/01 11:25 AM Page 397
Mappa Mundi (http://www.mappa.mundi.net/) a smart, monthly web-only

magazine and another Malamud/Webchick production, is perhaps the most
intellectual of the noncommercial online ‘zines.
Spark Online (http://www.spark-online.com/) is an extremely ambitious

monthly online magazine covering media, trends, and society. Like all the
others mentioned here, it is essentially a nonprofit labor of love.
The preceding sites show what can be done when original minds combine
fresh content with fine style.
Those directly following show what can be accomplished when innovation

and skillful graphic design are combined. Indeed, most of the following
sites exist solely for that purpose, though a few are also commercial in
nature.
Many of the sites listed require Flash and QuickTime, and it helps to have
a recent browser and a fast connection.
Amon Tobin Supermodified (http://www.amontobin.com/), previously men-

tioned in these pages, is an extraordinary music site created in Flash. A
cold, high-tech look, with a warmly interactive embrace, the site will
reward your patience.
Archinect (http://www.archinect.com/), an ever-changing visual explo-

ration, should be seen and not described.
Assembler (http://www.assembler.org/), Brent Gustaffson’s masterpiece of

cross-browser DHTML programming has a lovely and understated design
sensibility.
Born Magazine (http://www.bornmagazine.com/) is a long-running, ambi-

tious, collaborative work that attempts to continually reinvent the con-
junction between word and image. The noncommercial site’s tagline is
“Design. Literature. Together.”
Egomedia (http://www.egomedia.com/) is a design company portfolio with

the sensibility of a rock video. Requires Flash.
17 0732 CH13 4/24/01 11:25 AM Page 398
Lushly designed eneri.net (http://www.eneri.net/) makes no bones about

narrowing its audience: “This site targets luxurious people with a fast com-
puter, fast Internet connection, Netscape or IE 4.0 or above, and Shockwave
7 plug-in.” For those who meet the requirements, Irene Chan’s labor of love
offers a beautiful, film-like experience.
Entropy 8 Zuper (http://entropy8zuper.org/) is the site of Auriea Harvey, one

of the first web designers to laugh at conventions and bust boundaries.
Requires “fast computer, DHTML browser, Flash 5 or better” and “a physi-
cal need for wonder and poetry.”
Futurefarmers (http://www.futurefarmers.com/), Amy Franceschini’s web

and multimedia design company, gives the lie to the notion that corporate
work must be staid and conservative. Amy is one of the original exponents
of fine design on the Web; her early web work is housed permanently at
the San Francisco Museum of Modern Art.
Gmunk (http://www.gmunk.com/), a high-density personal site, pushes the

envelope every which way. Outrageously high bandwidth, QuickTime
movies, layered Photoshop collages rendered in Flash: everything usability
experts rail against is practiced here, by a master who can get away with
it. Tune in after losing an argument with your information designer or your
client.
Interiors (http://www.webproductions.com/photo/) is a dynamic slideshow

of digital self-portraits by artist Steve Giovinco. To call the work “disturb-
ing” would be an understatement. It’s also quite powerful.
Monocrafts (http://www.yugop.com/) combines powerful visual content

with unbelievably innovative interface ideas—extremely inspiring.
One9ine (http://www.one9ine.com/), a web design agency created by

designers, not marketers, is gorgeously rich yet entirely functional and easy
to navigate. Think the two can’t coexist? Look and see.
Once Upon A Forest (http://www.once-upon-a-forest.com/) is an abstract,

deliberately cryptic work of genius by Joshua Davis, who also brings us
Praystation and Dreamless.
17 0732 CH13 4/24/01 11:25 AM Page 399
S.M. Moalie’s Photomontage (http://www.photomontage.com/) makes us

cry. ‘Nuff said.
Marc Klein’s Pixel Industries (http://www.pixel-industries.com/), well

known and widely imitated, is a textbook example of the graphic-design-
lead approach to web development. See also Marc’s Creative Republic
(http://www.creative-republic.com/).
Pixelflo (http://pixelflo.com/), funky and witty, is also a masterpiece of

JavaScript programming.
Praystation (http://www.praystation.com/), Joshua Davis’s site, is dedicated

to exploring and enlarging the boundaries of Flash and interface design. If
you are learning Flash and beginning to think you know what it can and
can’t do, check Praystation. Davis gives away his source code so others can
use it in their design projects.
Presstube (http://www.presstube.com/), James Patterson’s personal illus-

tration portfolio, reveals mastery of Flash as well as considerable drawing
skill.
Projectbox (http://www.projectbox.com/) is an unusually elaborate, strik-

ingly designed illustration and design portfolio site made in Thailand by 22-
year-old Krisakorn Tantitemit. The playful and well-crafted interface makes
great use of frames and scripting, and the color combinations are uniquely
dramatic and pleasing.
Josh Ulm’s collaborative Remedi Project (http://www.theremediproject.

com/) is a bleeding-edge leading light. Requires a modern browser, a fist-
ful of plug-ins, and a fast connection (or great patience).
Mike Cina’s Trueistrue (http://www.trueistrue.com/) is a completely unique,

ever-changing, strangely minimalist exploration of line and form.
Volumeone (http://www.volumeone.com/) is Matt Owens’s masterpiece.

Updated quarterly, the site explores abstract visual issues through Flash
and Photoshop.
17 0732 CH13 4/24/01 11:25 AM Page 400
Yenz: The Secret Garden of Mutabor (http://www.yenz.com/) is a navigable

space of large, striking images that load quickly because they are entirely
vector-based. Created in Illustrator, Freehand, and Flash 3, the site guides
you through one rich image field after another. The effect is both mesmer-
izing and soothing.
ZX26 (http://www.zx26.com/) is a noncommercial Japanese font site, built

entirely with tiny animated GIFs and JavaScript.
Still hungry? The following design community sites showcase some of the
newest and funkiest work being done anywhere. In addition, most are lov-
ingly designed and cunningly programmed.
Design is Kinky (http://www.designiskinky.net/), created by Aussies Andrew

Johnstone and Jade Palmer, features designer mug shots and hosts inter-
views with the likes of David Carson (not that there are really any likes
of David Carson). For similar material, see Australia In Front (http://
www.australianinfront.com.au/).
Kaliber 10000 (http://www.k10k.net/), created by Danish lads Michael

Schmidt and Token Nygaard, publishes a special new design project every
week and is a superbly designed site in its own right—as you probably gath-
ered from the many times we’ve mentioned in this book.
Netdiver Net (http://www.netdiver.net/) feeds your eyes with links and your
brain with close-up interviews. Got a great site? Netdiver might review it
if it meets “chief imagineer” Carole Guevin’s criteria: The ‘diver seeks
impeccable content as well as superb design.
Japan-based Shift (http://www.shift.jp.org/) , the mother of all design por-

tals, has inspired most of the sites in this section. In addition to its online
presence, the site generates real-world design products such as the Gas-
book series and the IMG SRC 100 book.
Straight outta Luxembourg, Surfstation (http://www.surfstation.lu/) cur-

rently features the tiniest type on the Web. Fortunately the site’s design
news, interviews, and playful collaborative sections are easy and delight-
ful to read.
17 0732 CH13 4/24/01 11:25 AM Page 401
Three.oh Inspirational Kingdom (http://www.threeoh.com/) brags precision

design and special interactive features, as well as advanced and super-
funky JavaScript tricks. (For instance, loading an interview or special design
feature in a pop-up window causes the original window to be “grayed out”
by means of a full-screen layer swap.)
Special mentions:
Joe Jenett’s Coolstop (http://www.coolstop.com/v4.5/) was an independent

portal to fine design and original content long before there was even a cat-
egory for such sites, and continues to fulfill its mission with clarity, focus,
and integrity. Its spiritual predecessor was Glenn Davis and Teresa Martin’s
Project Cool, still operational, but not the same since its founders departed.
Notice that nearly every site mentioned is a noncommercial, independent

site. Coincidence? Read the Time Life Books.
Believe us when we tell you that the sites listed above are not even a frac-
tion of one percent of the best such sites out there. And there is always
room for more, which brings us to our valedictory address:
THE INDEPENDENT CONTENT PRODUCER

REFUSES TO DIE!
This book is written for professionals in a competitive market. Conse-
quently, we’ve spent most of our time talking about job skills—present and
future. But designers do not live by bread alone—not even when it’s really
good bread.
We’ve said it before and we’ll say it again: If the Web is fascinating simply
as a medium rife with challenges and rich in possibilities, it is even more
alluring when you consider its low barrier to entry. This medium does not
merely permit you to publish your own work, it begs for it.
17 0732 CH13 4/24/01 11:25 AM Page 402
402 HOW: Never Can Say Goodbye: The Independent Content Producer Refuses to Die!
From a purely selfish point of view, most of today’s best-known web

designers are famous for their personal sites, not for their commercial proj-
ects (though these are of course viewed and respected). Fame may seem a
silly thing to seek, but it sure doesn’t hurt when you’re looking for your next
job or your next client or approaching a backer to start your own agency.
The real jazz cats might do studio gigs to put three squares on the table,
but dawn always caught them blowing mad bop in crazy uptown clubs.
Real web designers jam after hours too—on personal and collaborative con-
tent and design sites, online magazines, and experimental spaces.
By creating and maintaining sites that cannot be controlled, compromised,

disfigured, or deleted by the indifference or poor judgement of clients or
managers, you will always have good work to show for yourself. More
importantly, you have the chance to express yourself—to find out what
you’re made of when no client is paying you and to find out what you really
want to say.
If you were a classical composer, you’d have to pay a symphony orchestra

just to hear your own music. And if you were a filmmaker, forget about it.
But in independent web production, the only questionable part of your
budget is how much time you can afford.
No one is in control of this space. No one can tell you how to design it, how
much to design it, when to “dial it down.” No one will hold your hand and
structure it for you. No one will create the content for you. What is in you?
What thwarted creative potential is burning to get out, grow, and find its
audience?
If you do this well, it will reflect back into the work you do for clients. Not
only will this help your career, it will also enrich your life and the lives of
others. Creating your content, designing it your way, repositioning yourself
from vendor to author, you will have made your mark on the medium and
perhaps on your generation.
You will have taken your talent to the Web.
18 0732 index 4/24/01 1:04 PM Page 403
Index
Symbols advantages of Java, 349-350

agnosticism, Web, 23-28
5k Contest, 50
AIGA (American Institute of Graphic Arts), 395
8-bit color, 57-59
algorithms
16-bit color, 57-59 GIF, 222
18-month pregnancy, 31 LZW, 234-236
aligning
elements, 271-272
A fonts, 240
images, 273
A List Apart, 182, 312, 390
Allaire Spectra, 167
absolute links, 186
ALT attribute, 90
absolute size keywords, 280-281
Amaya, 25
access
bandwidth, 41 Amazon.com, 77, 119
multimedia speed, 38-40 America Online (AOL), 115
accessibility, 65, 67 American Institute of Graphic Arts (AIGA), 395
Braille, 352-353 Amon Tobin Supermodified, 397
role of web designers, 145
analysis, phases of web projects, 152-156
text, 246
Andreessen, Marc, 116
active links, design conventions, 19-20
animation, GIF, 212, 223, 237-238
Active Server Pages (ASP), 333-339
anti-aliasing
ActiveX, 27
options, 243-244
activities, human activity on the Web, 35 troubleshooting, 244
add-ons, 360. See also plug-ins type, 241-243
adding inline styles, 269-270 antitrust lawsuit against Microsoft, 121
addresses Anuff, Joey, 118
FTP, 205 AOL (America Online), 115, 121
URLs, 180
Apple Computer, 113
Adobe, 395
Apple Internet Developer, 393
Illustrator, logos, 240
InDesign, 28 Apple Macintosh, 114
Advanced Research Projects Agency Network applets, embedding, 346
(ARPANET), 112-113
18 0732 index 4/24/01 1:04 PM Page 404
404 applications
applications authoring Flash, 372

FTP, 205-206 automated scripting, 250
Java, 344
avoiding progressive GIFs, 225
middleware, 332-335
Photoshop, 209-211
WYSIWYG, 202-204
B
applying
Flash, 373 B2B (Business-to-Business), 124, 139-140
JavaScript, 288-290 B2C (business-to-consumer), 140
browser detection/redirection, 312-315
Babble List, The, 390
default status, 299
event handlers, 295-298 Babelfish, 35
executing, 299-300 backgrounds, 212
global documents, 321-324 bandwidth, 14, 91
image rollovers, 302-306 caches, 49-50
links, 300-301 conserving, 44-48
platform/browser detection code, 316-320 CSS, 259
pop-up windows, 307-310 JavaScript, 26
pull-down menus, 310-312 LZW compression, 235
resources, 291-293, 324 traffic, 44
text rollovers, 294-295
LZW compression, 234-236 Bare Bones Guide to HTML, The, 183
sans serif fonts, 244 Barney’s, 106-107
SSI, 341-343 baseline styles, 240
style sheets, 267-270
Batman Forever, 117
Archinect, 397
BBEdit, 242
architecture, 32, 81
Beatnik, 368
ARPANET (Advanced Research Projects Agency),
Bell, Alexander Graham, 112
112-113
benefits of CSS (Cascading Style Sheets),
artists, web, 17
259-260
Ask Doctor Web, 181
Berners-Lee, Tim, 14, 115
ASP (Active Server Pages), 333-339
Bina, Eric, 116
Assembler, 397
Bloomingdales.com, 101
Astounding Websites, 390
blurring images, 231-233
Atlantic cable, 111
body text. See also text
attributes images, 36
ALT, 90 typography, 239
TITLE, 90
Born Magazine, 397
audio, 38
Boxtop Software website, 229
bandwidth, 41
Beatnik, 368 Braille, 352-353
MIDI, 368 brainstorming design phases of web projects, 156
quality issues, 38-40
WMP, 367
18 0732 index 4/24/01 1:04 PM Page 405
clarity 405
branding, 87, 103-107 text rollovers, 294-295

clarity, 89 multimedia, 352
IBM, 105 SMIL, 352-354
interfaces, 104 SVG, 354-358
Kioken Inc., 106 nongraphical, 259
role of web designers, 141-142 plug-ins, 358-362
Bray, Tim, 115 Beatnik, 368
developing content, 376-381
breadcrumbs, letting users know where they
QuickTime, 364-367
are, 101
RealPlayer, 362-364
breaking text into sub-units of information, Shockwave/Flash, 369-376
91-93 troubleshooting, 381-385
brevity, web site design, 90-91 WMP, 367
Browser Upgrade campaign, 389 sniffing, 305
source code, 185
browsers, 6, 253 upgrades, 31
absolute size keywords, 280 VRML, 350-351
caches, 49-50
compatibility, 204 Builder, 393
competition, 25 Builder.com, 182
CSS, 257 Bush, Vannevar, 112
compatibility, 261-262
Business-to-Business (B2B), 124, 139-140
content/style, 258, 261
design methods, 258-259 business-to-consumer (B2C), 140
modification benefits, 259-260 bytecode, 344
sizing fonts, 276-284
strategies, 274-275
style sheets, 263-270 C
troubleshooting, 271-273
elements, 271-272 caffemocha.com, 103
guidelines, 27 Cailliau, R., 115
GUIs, 85-87 Caillou, Robert, 14
HTML
calculating 16-bit colors, 58
capabilities, 254-256
CSS, 256 Cascading Style Sheets. See CSS
JavaScript, 288-290 case studies
default status, 299 Metafilter website, 337-338
detection/redirection, 312-315 Waferbaby website, 336
event handlers, 295-298 case-sensitivity, 180
executing, 299-300
global documents, 321-324 CERN, 115
image rollovers, 302-306 characters, editing, 240. See also fonts; typeface
links, 300-301 charging for plug-ins, 383
platform detection code, 316-320
chunking text, 93
pop-up windows, 307-310
pull-down menus, 310-312 clarity
resources, 291-293, 324 branding, 89
websites, 87
18 0732 index 4/24/01 1:04 PM Page 406
406 clarity
brevity, 90-91 Color Cube, 56, 212

icons, 88-90 colorblindness, 217
structural labels, 90 dithering, 213-218
text, 91-94 recurring hexadecimal pairs, 216
Clark, Jim, 116 web-safe color palettes, 215-219
clicking, 95-98 Color Palette dialog box, 219
client-side programming languages, 287 Color Picker dialog box, 219
client-side technologies, 330-331 colorblindness, 217
clients colors
design approval, 162 16-bit, 57-59
providing training for, 169-170 anti-aliasing, 241-243
selling ideas to, 158-160 comps, 160-164
dithering, 57
Cloninger, Curt, 17
gamma, 59-61
closing HTML tags, 177 Netscape Color Cube, 212
CNET.com, 77 colorblindness, 217
code dithering, 213-218
ASP, 335-339 recurring hexadecimal pairs, 216
browser/platform detection (JavaScript), web-safe color palettes, 215-219
316-320 palettes, 210, 219-221
bytecode, 344 viewing, 55-57
colors, 56 web-safe fonts, 245
HTML commands
browser compatibility, 204 File menu
closing tags, 177 Save For Web, 229
comments, 200-201 View Page Info, 184
constructing tags, 178 View menu, Source, 183
conventions, 176-177 COMMENT tags, 200-201
formatting, 179-181
commerce, 22
frames, 194-195
links, 185-187 commercial interactivity, 22
META tags, 197-200 communicating
plug-ins, 193-195 functionality of web sites, 164
sensible markup, 189 through web design, 137-138
tables, 193-195 through web sites, 140-141
tools, 190-192 Communication Arts, 395
tutorials, 181-182
communities, online, 35
URLs, 180-181
validating, 188 comparing print to Web rich media, 327-329
viewing, 183-185 compatibility
WYSIWYG applications, 202-204 browsers, 204
PHP, 335-339 CSS, 261-262
source code, 42-46 users’ needs, 145
Cold Fusion, 334-338 compensating for gamma issues, 60, 211
competition, browsers, 25
18 0732 index 4/24/01 1:04 PM Page 407
conventions 407
composition typography, 239-240

creating color comps, 163-164 CSS , 256
design, 160-162 HTML, 254-256
images, 210 troubleshooting, 244-246
presenting, 161-162 web pages
slicing, 211 ASP/PHP, 335-339
compression middleware applications, 334-335
ImageReady, 230 server-side technologies, 331-334
JPEG, 231-233 websites, 387-388
lossless, 222 connections
lossy, 226 bandwidth, 41
LZW, 234-236 multimedia, 38-40
Computer Currents High-Tech Dictionary, 129 conserving bandwidth, 44-50
conditions consistency guidelines for interfaces, 82-83
creating, 331-334 constructing HTML tags, 178
middleware applications, 334-335
containers, content, 342
configuring
content, 70
CSS, 257
characteristics, 22
compatibility, 261-262
containers, 342
content/style, 258-261
CSS, 258-261
design methods, 258-259
plug-ins, 376-381
modification benefits, 259-260
presentation, 274-275
strategies, 274-275 continuing education, 387-389
style sheets, 263-270 A List Apart, 390
troubleshooting, 271-273 Astounding Web sites, 390
external style sheets, 267-268 Babble List, The, 390
gamma, 59-61 Dreamless, 391
images, 221 Evolt, 391
animated GIFs, 237 Metafilter, 391
compressing, 231-233 online classes, 392
creating animated GIFs, 238 Redcricket, 392
GIF, 222-225 viewing websites for ideas, 393-401
JPEG, 226, 228 Webdesign-1, 392
optimizing, 228-230 conventions
PNG, 236-237 browsers, 27
logos, 356 HTML, 176-177
navigation, 246-247 closing tags, 177
Netscape Color Cube, 212 constructing tags, 178
colorblindness, 217 formatting, 179-181
dithering, 213-218 URLs, 180-181
recurring hexadecimal pairs, 216 IIS, 181
web-safe color palettes, 215-219 multimedia, 330, 352
rollovers, 293 SMIL, 352-354
semantic websites, 251-252 SVG, 354-358
Web, 20
18 0732 index 4/24/01 1:04 PM Page 408
408 converting RGB (Red, Green, and Blue)
converting RGB (Red, Green, and Blue), 240 deployment, 166

Cool Site of the Day, 117 learning client’s methods, 169-170
providing client training, 169-170
copying interfaces from other sites, 77-78
providing documentation and style guides, 168
creating updating, 167-168
color comps, 163-164
design. See also formatting
effective sites, 71-72
architecture, 32
style guides, 168
browser, 18-month pregnancy, 31
Creating Killer Websites, 221 conventions, 20
crisp anti-aliasing, 241. See also anti-aliasing HTML, 175-176
Crozier, Steve, 1 browser compatibility, 204
closing tags, 177
CSS (Cascading Style Sheets), 256-257, 335 code conventions, 176-177
absolute size keywords, 280-281 comments, 200-201
compatibility, 261-262 constructing tags, 178
content/style, 258, 261 formatting, 179-181
design methods, 258-259 frames, 194-195
inline styles, 269-270 links, 185-187
modification benefits, 259-260 META tags, 197-200
relative keywords, 281-282 plug-ins, 193-195
style sheets, 263-270 sensible markup, 189
sizing fonts, 276-284 tables, 193-195
strategies, 274-275 tools, 190-192
troubleshooting, 271-273 tutorials, 181-182
CSS-1 standard, 261 URLs, 180-181
customizing anti-aliasing, 243-244 validating, 188
viewing source code, 183-185
CuteFTP, 205
WYSISWYG applications, 202-204
interfaces, 72-74, 80-81
D Liquid Design, 51-55
for non-traditional devices, 97
databases, 333 opportunities, 401-402
overview, 23
Dave Raggett’s “Getting Started With HTML,” 182
phases of web projects, 156
David Siegel’s website, 221 approval by client, 162
Davis, Glenn, 117 brainstorming and problem solving, 156
DeBabelizer, 238 color comps, 160-162
requirements, 157
debugging relative links, 186-187
selling ideas to clients, 158-160
defaults, status (JavaScript), 299 tools, 65
defining viewer control, 36
content characteristics, 22 Web
HEAD tag, 308-310 agnosticism, 23-26
deleting redundancy, 47-48 for an ever changing Web, 9-12
web pages, 20-23
Design is Kinky, 400
18 0732 index 4/24/01 1:04 PM Page 409
ECMAScript, 409
design technicians, 131-132 comments, 200-201

Designing With JavaScript: Creating Dynamic Web constructing tags, 178
Pages, 291 CSS, 256
cutting/pasting, 339-340
desktop publishing, 114
embedding, 268-269, 346, 376-378
detecting formatting, 179-181
browsers, 312-315 frames, 194-195
platforms, 316-320 HEAD tag, 308-310
deuteranopia (colorblindness), 217 links, 185-187
development META tags, 197-200
phases of web projects, 162-163 plug-ins, 193-195
creating color comps, 163-164 sensible markup, 189
designing for easy maintenance, 165-166 tables, 193-195
functionality, 164 tools, 190-192
working with templates, 165 tutorials, 181-182
plug-in content, 376-381 typography, 254-256
URLs, 180-181
device-independence, open standards, 23-29 validating, 188
DHTML (Dynamic HTML), 21, 334-335 viewing source code, 183-185
dial-in connections, bandwidth, 41 WYSIWYG applications, 202-204
HTML, 332
dialog boxes
Color Palette, 219
Color Picker, 219 Dolby, Thomas, 368
Matte Color, 228 DOM (Document Object Model), 128, 285
Photoshop, Slices, 249 domain name servers (DNS), 113
Save For Web (Photoshop), 229
Swatches, 219 dot.coms, failure of, 80
Type Tool (Photoshop), 241 downloading, brevity of downloads, 90-91
dicing images, 211 Dreamless, 36, 391
Digital Subscriber line (DSL), 41 Dreamweaver, 28, 46
Digital Web, 83, 393 drymartini.com, 89
disabilities, people with, 246 DSL (Digital Subscriber Line), 41
disappearing websites, 6 Dynamic HTML (DHTML), 21, 334-335
dithering, 57, 213-218 dynamic websites, 329-331
DNS (domain name servers), 113
Document Object Model (DOM), 128, 285 E
documents, 285. See also web pages
global (JavaScript), 321-324 e-commerce, 22, 119-120
HTML, 175-176 ease of learning, guidelines for interfaces, 82
adding inline styles, 269-270 ECMA (European Computer Manufacturers
applying SSIs, 341-343 Association), 27, 288
browser compatibility, 204
closing tags, 177 ECMA-62, 288
code conventions, 176-177 ECMAScript, 26, 127, 288
18 0732 index 4/24/01 1:26 PM Page 410
410 editors.
editors. See also applications handlers, 285, 295-298

fonts, 240 history of, 287
Photoshop, 209-211 image rollovers, 302-306
text, 184-185 links, 300-301
WYSIWYG, 202-204 platform/browser detection code, 316-320
education, 389 pop-up windows, 307-310
A List Apart, 390 pull-down menus, 310-312
Astounding Websites, 390 resources, 291-293, 324
Babble List, The, 390 text rollovers, 294-295
Dreamless, 391 Evolt, 391
Evolt, 391 executing
Metafilter, 391 JavaScript, 299-301
online classes, 392 Visibone color palette, 219, 221
Redcricket, 392
exporting Web formats, 210-211
viewing websites for ideas, 393-401
Webdesign-1, 392 Extensible Hypertext Markup Language (XHTML),
29, 128
effective sites, creating, 71-72
Extensible Markup Language (XML), 125
Egomedia, 397
external style sheets, 267-268
eight-bit (256 color) systems, Netscape Color
Cube, 213 extranets, 124-125
elements, positioning, 271-272
embedding F
applets, 346
Flash, 376 failure of dot.coms, 80
plug-ins, 377-378 feathering images, 231
Style Sheets, 268-269
feedback, guidelines for interfaces, 84
Entropy 8 Zuper, 398
File menu commands
environments Save For Web, 229
Java View Page Info, 184
advantages of, 349-350
File transfer Protocol (FTP), 205-206
Java Virtual Machine, 344-346 files
CSS, 268
European Computer Manufacturers Association
streaming (JavaScript), 307-310
(ECMA), 27, 288
WMF (Windows Media File), 367
events
filling, anti-aliasing, 241-243
JavaScript, 285-286
applying, 288-290 Filo, David, 117
browser detection/redirection, 312-315 Flash, 369-372, 374-376
default status, 299 embedding, 376
event handlers, 295-298 logos, 356
executing, 299-300 plug-in, 349
global documents, 321-324 Flash to the Core, 361
18 0732 index 4/24/01 1:04 PM Page 411
Graphical User Interface (GUI) 411
Flash Web Design: The Art of Motion HTML, 254-256

Graphics, 361 troubleshooting, 244-246
Fleming, Jennifer, 82 web pages
ASP/PHP, 335, 337-339
Flution 1.5, 354
fonts, 62-64 server-side technologies, 331-334
CSS, 276-284 websites, 387-388
horizontal/vertical type, 240-243
frames, 26, 96, 194-195, 334-335
HTML, 254
inline styles, 269-270 framesets, 26, 195
spacing, 245 Fray, 36, 396
troubleshooting, 244-246 Freehand logos, 240
web-safe colors, 245
FTP (File Transfer Protocol), 205-206
formats, exporting, 210-211
functional spec, 334
formatting
functionality, phases of web projects, 164
anti-aliasing, 243-244
CSS, 257 Furbo Filters Webmaster, 60
compatibility, 261-262 Futurefarmers, 398
modification benefits, 259-260 G
strategies, 274-275 gamma, 59-61, 211
style sheets, 263-270 GammaToggle FKEY, 60
troubleshooting, 271-273 Gap, 76
external style sheets, 267-268
gamma, 59-61 Gates, Bill, 119
HTML, 179-181 generating tweens, 239
images, 221 GIF (Graphics Interchange Format), 222-225
animated GIFs, 237 animating, 212, 237
creating animated GIFs, 238 ImageReady, 238
GIF, 222-225 LZW compression, 235-236
JPEG, 226-228 optimizing, 228-230
optimizing, 228-233 transparent, 255
PNG, 236-237
Glassdog, 87, 396
logos, 356
navigation, 246-247 glish, 86
Netscape Color Cube, 212 global documents
colorblindness, 217 JavaScript, 321-324
dithering, 213-218 resources, 324
recurring hexadecimal pairs, 216 Gmunk, 398
web-safe color palettes, 215, 218-219
GoLive, 28, 46
rollovers, 293
semantic websites, 251-252 Google, 332
typography, 239-240 graphical devices, 88. See also icons
CSS, 256 Graphical User Interface (GUI), 84-85, 113, 211
18 0732 index 4/24/01 1:04 PM Page 412
412 graphics
graphics, 221. See also images HTML (Hypertext Markup Language), 125. See
animated GIFs, 237-238 also web pages
GIF, 222-230 BBEdit, 242
JPEG, 226-228 browser compatibility, 204
compressing, 231-233 code
optimizing, 228-230 conventions, 176-177
Netscape Color Cube viewing, 183-185
colorblindness, 217 cutting/pasting, 339-340
dithering, 213-218 defined, 175-176
recurring hexadecimal pairs, 216 documents, 256. See also web pages
web-safe color palettes, 215-219 Flash, 376
PNG, 236-237 formatting, 179-181
SVG, 354-358 frames, 194-195
Graphics Interchange Format. See GIF GIFs, 225
HEAD tag, 308
Grey Advertising, Inc., 123
history of, 14-17, 19, 25
GUI (Graphical User Interface), 84-85, 113, 211 images rollovers (JavaScript), 302-306
guidelines Java applets, 346
browsers, 27, 85-87 JavaScript
for interfaces, 82-84 executing, 299-300
links, 300-301
jockeys, 131-132
H links, 185-187
plug-ins, 193-195, 377-378
Happy Cog website, 315 practitioners, 131-132
Harrumph!, 92, 396 redundancy, 47-48
HEAD tag (HTML), 308 sensible markup, 189
SSI , 341-343
headlines, invisible, 255 style sheets, 268-269. See also CSS
hesketh.net, 100 tables, 176, 193-195, 257
hexadecimal pairs, recurring, 216 tags
adding inline styles, 269-270
hierarchies, 20-23, 97-98
closing, 177
High Five, 118 comments, 200-201
highlighting, letting users know where they are in constructing, 178
the site, 101-102 CSS, 256
histories META, 197-200
HTML, 25 typography, 254-256
JavaScript, 286-287 technicians, 131-132
Web, 14-17, 19 text rollovers (JavaScript), 294-295
tools, 190-192
history of the Web, 111-122
tutorials, 181-182
horizontal scrolling, 96 URLs, 180-181
horizontal type, inserting, 240-243 validating, 188
hosts, 114 viewing, 342
WYSIWYG applications, 202-204
Hot Sauce, 351
18 0732 index 4/24/01 1:04 PM Page 413
interfaces 413
HTML 4 Transitional Recommendation, 26 Netscape Color Cube, 212

HTML 4.01 Specification, 183 colorblindness, 217
dithering, 213-214, 218
HyperCard, 114
recurring hexadecimal pairs, 216
hyperlinks, 6, 19-20, 112, 185-187 web-safe color palettes, 215, 218-219
hypertext, 6, 112, 115 overlapping, 273
Hypertext Markup Language. See HTML Photoshop, 209-211
PNG, 236-237
rollovers, 302-306
I sharpening, 231-233
slicing, 248-250
Ian S. Graham’s “Introduction to HTML,” 181 SVG, 354-358
IBM (International Business Machine), 105, 142 swapping, 211
tiling, 212
iCab, 25
ImageVice, 229
Icon Factory, 78
implementation
icons CSS
clarity, 88-90 compatibility, 261-262
invisible text labels, 90 sizing fonts, 276-284
ideas, selling to clients, 158-160 strategies, 274-275
IIS (Internet Information Server), 181 style sheets, 263-270
illustrations, 223. See also GIF
interactivity, 23
Illustrator logos, 240
impossible agreements, 158
image swap, 84
includes, 341
ImageReady (JavaScript), 46, 250
Information Superhighway, 114
animated GIFs, 238
compression, 230 inserting inline styles, 269-270
images, 221 Integrated Digital Services Network (ISDN), 45
aligning, 273 interactive behaviors
animated GIFs, 237 continuing education, 387-388
anti-aliasing, 243-244 interfaces, 78-79
blurring, 231-233 interactivity, 20-23, 65
body text, 36 client-side/server side technologies, 330-331
creating animated GIFs, 238 applying SSIs, 341-343
dicing, 211, 248-250 cutting and pasting, 339-340
feathering, 231 JavaScript, 286
GIF, 222-225
animation, 212 interfaces. See also browsers
optimizing, 228-230 architecture, 81
horizontal/vertical type, 240-243 branding, 104
JPEG, 226, 228 colors
compressing, 231-233 16-bit, 57-59
optimizing, 228-230 gamma, 59-61
LZW compression, 234-236 viewing, 55-57
moving, 21 copying from other sites, 77-78
design of, 80-81
18 0732 index 4/24/01 1:04 PM Page 414
414 interfaces
formatting, 72-74 J
GUI (Graphical User Interface), 84-87, 211
guidelines for, 82 Java, 343-344
consistency, 82-83 advantages of, 349-350
continual feedback, 84 applets, 346
ease of learning, 82 HTML, 346
interactive behaviors, 78-79 troubleshooting, 347-348
Liquid Design, 51-55 Java Virtual Machine, 344-346
navigation
JavaScript, 26, 119, 127, 285-286, 330
elements, 74-76
applying, 288-290
purpose for, 80
default status, 299
sizing, 51
event handlers, 295-298
Interiors, 398 executing, 299-300
Internet, 6, 125-126 global documents, 321-324
accessibility, 65, 67 history of, 287
communities, 35 image rollovers, 302-306
growth of, 22 ImageReady, 250
multimedia, 37-40 links, 300-301
standards, 25 platform/browser detection code, 316-320
typography, 62-64 pop-up windows, 307-310
Internet Channel, 45 pull-down menus, 310-312
resources, 291-293, 324
Internet Explorer, 119-120. See also browsers
rollovers, 225
18-month pregnancy, 31
absolute size keywords, 280-281
Macintosh Edition, 121 JavaScript for the World Wide Web: Visual
source code, 185 Quickstart Guide, Third Edition, 292
Internet Information Server (IIS), 181 Jay Boersma’s Web Work, 181
Internet Protocol (IP), 125 jazzradio.net, 102
Internet Relay Chat (IRC), 114 Jobs, Steve, 45, 113
Internet Service Provider (ISP), 45 Joe Jenett’s Coolstop, 401
intranet, 126-127 JPEG (Joint Photographic Experts Group),
226-228
invisible headlines, 255
compressing, 231, 233
invisible text labels, 90 optimizing, 228-230
IP (Internet Protocol), 125 JScript, 287
IpixViewer plug-in, 346 Justin’s Links from the Underground, 117
IRC (Internet Relay Chat), 114
ISDN (Integrated Digital Services Network), 45
ISP (Internet Service Provider), 45
18 0732 index 4/24/01 1:04 PM Page 415
Marc Klein’s Creative Republic 415
K image rollovers (JavaScript), 302-306

JavaScript
Kaliber 10000, 400 default status, 299
kerning, 240 event handlers, 295-298
executing, 299-301
keyboard shortcuts (Photoshop), 241
platform/browser detection code (JavaScript),
keywords 316-320
absolute size, 280-281 pop-up windows (JavaScript), 307-310
relative, 281-282 pull-down menus (JavaScript), 310-312
Kioken Inc., branding, 106 resources (JavaScript), 324
text rollovers (JavaScript), 294-295
Linux, 180
L Liquid Design, 51-55
labels, 90 listening to clients, 137
Lance Arthur’s Design-o-Rama, 182 LiveMotion, 371
languages LiveScript, 26, 287. See also JavaScript
JavaScript, 286 locations
Perl, 286 letting users know where they are in the site,
SMIL, 352-354 101-102
leading, 240, 273 window.location (JavaScript), 315
learning Web standards, 29-30 logos
left-aligned images, 273 GIF, 223
SVG, 356, 358
left-hand navigation bars, 77-78
typography, 240
length units, 282-283
lossless compression, 222
Licklider, Dr J.C.R., 112
lossy compression, 226
life cycles of projects, 148
Lynda Weinman’s website, 221
limitations of mediums, 142-145
Lynx, 25
accessibility, 145
compatibility with users’ needs, 145 LZW compression, 234-236
team work, 144
technology, 143
visually appealing, 144-145 M
line art, 223. See also GIF Macintosh, 113
line-height property (CSS), 273 source code, 184
links, 19, 185-187. See also hyperlinks sRGB (standard RGB), 61
browser detection/redirection (JavaScript), typographic preferences, 63-64, 277-279
312-315 Macromedia Freehand, 240
CSS files, 268 maintaining web sites, 165-166
design conventions, 20
Mappa Mundi, 397
global documents (JavaScript), 321-324
Marc Klein’s Creative Republic, 195
18 0732 index 4/24/01 1:04 PM Page 416
416 Marc Klein’s Pixel Industries
Marc Klein’s Pixel Industries, 399 movies, 38-40

markup, 342. See also HTML bandwidth, 41
plug-ins, 377-378 pop-up windows JavaScript, 307-310
redundant, 339-340 QuickTime, 364-367
sensible, 189 RealPlayer, 362-364
WMP, 367
Mary Quant site, 74
moving images, 21
Matte Color dialog box, 228
Mozilla, 121
media, rich. See multimedia; rich media
multimedia, 37, 352
medium restrictions, 142-145
bandwidth, 41
accessibility, 145
interactivity, 22
compatibility with users’ needs, 145
plug-ins, 358-362
team work, 144
Beatnik, 368
technology, 143
developing content, 376-381
visually appealing, 144-145
QuickTime, 364-367
menus, pull-down (JavaScript), 310-312 RealPlayer, 362-364
META tags, 197-200 Shockwave/Flash, 369-376
meta-languages, 178 troubleshooting, 381-385
WMP, 367
Metafilter website, 337-338 pop-up windows (JavaScript), 307-310
methodologies, 149, 151 quality issues, 38-40
Microsoft SMIL, 352-354
antitrust lawsuit, 121 SVG, 354-358
Internet Explorer, 25, 31, 119 web standards, 330
JScript, 287 multiple web pages, external style sheets,
middleware applications, 332-335 267-268
MIDI (Musical Instrument Digital Interface), 368 Musical Instrument Digital Interface (MIDI), 368
Mike Cina’s Trueistrue, 399 MySQL, 333
minimizing bandwidth, 49-50
missing plug-ins, replacing, 346 N
modems, bandwidth, 41
Narcotics Anonymous, 69-72
modifying. See also formatting
CSS, 259-260 National Center for Supercomputing Applications
default status (JavaScript), 299 (NCSA), 116
images, 240-243 navigation
interfaces, 51 bars, 77-78
monitors, Liquid Design, 51 devices, 102-103
formatting, 246-247
Monocrafts, 398
interfaces, 74-76, 81
Mosaic, 25, 116 slicing/dicing, 248-250
Mosaic Communications Corporation, 116 web pages, 20-23
mouse event handlers, 296-298 websites, 19
18 0732 index 4/24/01 1:04 PM Page 417
phases of web projects 417
NCSA (National Center for Supercomputing Open Text, 115

Applications), 116 Opera, 25, 122
Nelson, Ted, 112 opportunities on the Web, 401-402
netadmins (network administrators), 131 optimize palette, 239
Netdiver Net, 400 optimizing
Netscape, 6, 25, 116-117, 121. See also browsers GIF, 228-230
18-month pregnancy, 31 JPEG, 228-230
Beatnik plug-in, 368 options, anti-aliasing, 243-244
Color Cube, 56, 212
overlapping images, 273
colorblindness, 217
dithering, 213-214, 218 overstock.com, 79
recurring hexadecimal pairs, 216 overview of web design, 23
web-safe color palettes, 215, 218-219
JavaScript. See JavaScript
source code, 184-185 P
network administrators, 131
palettes
NeXT, 24, 115 colors, 210
Nielsen, Jakob, 16 optimize, 239
non-animated GIFs, 223 Visibone, 219-221
web-safe colors, 215, 218-219
non-graphical browsers, 259
Palm Pilots, 29
non-traditional devices, designing for, 97
patterns, 212
nonstandard workarounds, 31
PDA (Personal Digital Assistant), 24, 259
NSFNET, 114
designing for, 97
icons, 90
O PDN-Pix, 396
percentage units, 283-284
object-oriented programming languages, 343.
Perl, 286, 330
See also Java
Personal Digital Assistant. See PDA
objects, 285, 344
Personal Home Page tool (PHP), 331
Once Upon A Forest, 398
One9ine, 398
analysis, 152-156
online classes, 392 deploying the site, 166
online communities, 35 learning client’s methods, 169-170
onLoad event handler (JavaScript), 298 providing client training, 169-170
providing documentation and style
onMouseOut event handler (JavaScript), 295-298
guides, 168
onMouseOver event handler (JavaScript), updating, 167-168
295-298 design, 156
open standards approval by client, 162
device-independence, 29 brainstorming and problem solving, 156
learning, 29-30 color comps, 160-162
platform-agnosticism, 27-28 requirements, 157
selling ideas to clients, 158-160
18 0732 index 4/24/01 1:04 PM Page 418
418 phases of web projects
development, 162-163 pop-up windows (JavaScript), 307-310

creating color comps, 163-164 Populi Curriculum in Web Communication
designing for easy maintenance, 165-166 Design, 2
functionality, 164
Populi program, 2
working with templates, 165
testing, 166 positioning elements, 271-272
PhotoGIF, 229 Pratt Institute, 2
Photomontage web site, 17 Praystation, 399
Photoshop, 56, 209-211 preparation
commands, Save For Web (File menu), 229 composition, 210
dialog boxes typography, 211
Slices, 249 presenting
Type Tool, 241 color comps, 161-162
GIFs, 224-225 separating content, 274-275
ImageReady, 239 Presstube, 399
images, 248-250
keyboard shortcuts, 241 Presstube website, 17
PHP (Personal Home Page tools), 331-339 printing
comparing to Web rich media, 327-329
Pixelflo, 399 web pages, 88
pixels, 278-279 printing press, invention of, 111
placement of navigational devices, 102-103 problem solving, phases of web projects, 156
platform-agnostic, 23-28 producers, 131-132
plug-ins, 350, 358-362 programmers, 129-130
Beatnik, 368
content, 376-381 programming languages
embedding, 377-378 ECMAScript, 288
Flash, 349 Java, 343
HTML, 193-195 JavaScript, 286
IpixViewer, 346 Perl, 286, 330
Java applets, 346 SMIL, 352-354
JavaScript, 316-320 progressive GIFs, 225
multimedia, 352 progressive JPEGs, 228
SMIL, 352-354
Project Cool’s Gettingstarted.net, 181
SVG, 354-358
QuickTime, 38, 346, 364-367 project managers, 130-131
RealPlayer, 362-364 Projectbox, 399
redirecting, 379-381 projects, life cycles of, 148
Shockwave/Flash, 369-376
properties, line-height (CSS), 273
VRML, 350-351 pseudo-science, 77
WMP, 367 publishing websites, 205-206
PNG, 236-237 pull-down menus (JavaScript), 310-312
points, 276-277
18 0732 index 4/24/01 1:04 PM Page 419
rules, style sheets 419
Q resources
JavaScript, 291-293, 324
QA (quality assurance), 97 websites, 389
Quark XPress, 28 A List Apart, 390
Astounding Websites, 390
QuickTime, 38, 346, 364-365, 367. See also
Babble List, The, 390
multimedia; video
Dreamless, 391
quotation marks (JavaScript), 299 Evolt, 391
Metafilter, 391
online classes, 392
R Redcricket, 392
viewing for ideas, 393-401
raster images, 221. See also images
Webdesign-1, 392
animated GIFs, 237-238
GIF (Graphics Interchange Format), 222-225, responsibilities of web designers, 136-138
228-230 B2B, 139-140
JPEG, 226-228 B2C, 140
compressing, 231-233 communication, 140-141
optimizing, 228-230 look and feel of web sites, 138-139
PNG, 236-237 medium restrictions, 142-145
reading on screen, 91-93 revisions, fonts, 240
RealPlayer, 42, 362-364 RGB (Red, Green, and Blue), 56, 240.
See also colors
RealSlideshow authoring tool, 353
rich media, 327-329, 350
recurring hexadecimal pairs, 216
multimedia, 352. See also multimedia
Red, Green, and Blue (RGB), 56, 240. SMIL, 352-354
See also colors SVG, 354-58
Redcricket, 36, 392 plug-ins, 358-362
redirecting Beatnik, 368
browsers, 312-315 developing content, 376-381
plug-ins, 379-381 QuickTime, 364-365, 367
RealPlayer, 362-364
redundancy Shockwave/Flash, 369-376
deleting, 47-48 troubleshooting, 381-385
replacing markup, 339-340 WMP, 367
referrer logs, 39 VRML, 350-351
relative keywords, 281-282 rollovers, 211, 294-295
relative links, 185-187 creating, 293
image, 302-306
reloading META tags, 200
JavaScript, 225
replacing
Ron Woodall’s HTML Compendium, 183
missing plug-ins, 346
redundant markup, 339-340 royalties, GIF (Graphics Interchange Format), 223
requirements for design phases of web “Rule of Five,” 99-100
projects, 157 rules, style sheets, 263-266
resolution, 55-61, 278-279. See also colors
18 0732 index 4/24/01 1:04 PM Page 420
420 S.M. Moalie’s Photomontage
S server-side technologies, 330-331

applying SSIs, 341-343
S.M. Moalie’s Photomontage, 399 cutting/pasting, 339-340
sans serif fonts, 240, 244. See also fonts Java, 343-344
advantages of, 349-350
Sapient.com, 94
Java Virtual Machine, 344-346
Save For Web command (File menu), troubleshooting, 347-348
Photoshop, 229 web pages
Scalable Vector Graphics (SVG), 354-358 creating, 331-334
screens. See interfaces middleware applications, 334-335
scripting servers, 26, 181

JavaScript, 285-286 servlets, 344, 349
applying, 288-290 SGI (Silicon Graphics Machines), 59
SGML (Standard Generalized Markup Language),
default status, 299
125, 178
event handlers, 295-298
executing, 299-300 sharpening images, 231-233
global documents, 321-324 Shift, 400
history of, 287 Shockwave, 369-376
image rollovers, 302-306
shortcuts, Photoshop, 241
links, 300-301
platform/browser detection code, 316-320 Siegel, David, 17, 117-118
pop-up windows, 307-310 Silicon Graphics Inc., 116
pull-down menus, 310-312 Silicon Graphics Machines (SGI), 59
resources, 291-293, 324
text rollovers, 294-295 site development, 6. See also websites
ImageReady, 250 content of, 70
life cycles of projects, 148
scrolling, 95-96
sizing. See also modifying
search engines, 115, 332 absolute size keywords (CSS), 280-281
searching fonts, 240, 276-284
META tags, 197-200 interfaces, 51
text, 246 relative keywords (CSS), 281-282
Section 508 of the Workforce Investment Act, 66 typography, 278-279
selecting typography, 239-240 Slices dialog box, Photoshop, 249
semantic websites, 251-252 slicing images, 211
semicolons (JavaScript), 300 SMIL (Synchronized Multimedia Integration
Language), 352-354
sensible markup, 189. See also HTML
smoothing type, anti-aliasing, 241-243
separation
content, 274-275 sniffing browsers, 305
content/style (CSS), 258, 261 sound, 38
serif fonts, 240. See also fonts bandwidth, 41
Beatnik, 368
Server Side Includes (SSI), 331
MIDI, 368
server-side scripting languages, Perl, 286 quality issues, 38-40
WMP, 367
18 0732 index 4/24/01 1:04 PM Page 421
team work, role of web designers 421
source code, 42-46, 183-185 Suck, 118-119

Source command (View menu), 183 Sun, 119
spacing fonts, 245 Surfstation, 400
Spark Online, 397 SVG (Scalable Vector Graphics), 354-358
specifying anti-aliasing, 243-244 swapping images, 211
spiders, 332 Swatches dialog box, 219
sRGB (standard RGB), 59-61. See also colors Synchronized Multimedia Integration Language
SSI (Server Side Includes), 331 (SMIL), 352-354
Standard Generalized Markup Language (SGML), syntax, JavaScript, 299-300
125, 178 sysadmins (systems administrators), 131
standard RGG (sRGB), 59
standards T
browsers, 27
multimedia, 330, 352 tables, HTML, 176, 193-195, 255
SMIL, 352-354
tags
SVG, 354-358
HTML, 175-176
open
adding inline styles, 269-270
device-independence, 29
browser compatibility, 204
learning, 29-30
closing tags, 177
platform-agnosticism, 27-28
code conventions, 176-177
Web, 25
comments, 200-201
statistics, Web users, 8 constructing tags, 178
status, default (JavaScript), 299 CSS, 256
Steadman, Carl, 118 embedding, 268-269, 346
formatting, 179-181
strategies, CSS (Cascading Style Sheets), 274-275 frames, 194-195
streaming video, 307-310. See also multimedia HEAD, 308-310
strength of Web, 32 links, 185-187
META, 197-200
structural labels, 90
plug-ins, 193-195
style guides, creating, 168 sensible markup, 189
style sheets tables, 193-195
CSS, 263-270 tools, 190-192
sizing fonts, 276-284 typography, 254-256
strategies, 274-275 URLs, 180-182
troubleshooting, 271-273 validating, 188
embedding, 268-269 viewing source code, 183-185
external, 267-268 WYSIWYG applications, 202-204
inline styles, 269-270 icons, 90
leading, 273 Taylor, Robert, 113
styles TCP (Transport Control Protocol), 125
CSS (Cascading Style Sheets), 258, 261
team work, role of web designers, 144
inline, 269-270
18 0732 index 4/24/01 1:04 PM Page 422
422 technologies
technologies traffic, bandwidth, 44-50

medium restrictions, 143 training clients, 169-170
standards, 29-30
transactions, 330-331
and the Web, 7-9
advantages of Java, 349-350
telegraph, invention of, 111 applying SSIs, 341-343
telephones, invention of, 112 creating web pages, 331-334
templates cutting/pasting, 339-340
database-driven websites, 332-334 Java, 343-344
working with, 165 Java Virtual Machine, 344-346
testing phases of web projects, 166
troubleshooting Java, 347-348
text. See also documents
Transport Control Protocol (TCP), 125
Braille, 352-353
chunking, 93 Transmit, 205
CSS, 256-257 transparent GIFs, 255
compatibility, 261-262 transparent images, 242
troubleshooting
anti-aliasing, 244
modification benefits, 259-260
CSS, 271-275, 276-284
Java, 347-348
strategies, 274-275
navigation, 246-247
style sheets, 263-270
plug-ins, 381-385
SVG, 357
editors, 184-185
typography, 244-246
fonts, 62-64
HTML, 254-256 tutorials
images, 36 HTML, 181-182
JavaScript, 299 JavaScript, 292, 324
labels, 90 tweens, generating, 239
reading on screen, 91-93 type
rollovers, 294-295 anti-aliasing, 241-244
searching, 246 horizontal/vertical, 240-243
zooming, 64
Type Tool dialog box (Photoshop), 241
“Three-Click Rule,” 97-98
types of style sheets, 267-270
Three.oh Inspirational Kingdom, 401
typography, 62-64, 93-94
tiling images, 212 CSS
TITLE attribute, 90 sizing fonts, 276-284
tools, 65 strategies, 274-275
BBEdit, 242 dithering, 213-214, 218
DeBabelizer, 238 formatting, 239-240
dicing, 248-250 GIF, 223
HTML, 190-192 HTML
inline styles, 270 CSS, 256-262
Photoshop, 209-211 tags, 254-256
slicing, 248-250 inline styles, 269-270
preferences, 277-279
tracking, 240
preparing, 211
18 0732 index 4/24/01 1:04 PM Page 423
Web 423
print, 327-329 QuickTime, 364-365, 367

sizing, 278-279 RealPlayer, 362-364
troubleshooting, 244-246 WMP, 367
web-safe colors, 245 View menu commands, Source, 183
View Page Info command (File menu), 184
U viewer control, 36
viewing
UDP (User Datagram Protocol), 125 colors, 55-57
Uniform Resource Locator (URL), 6, 180 16-bit, 57-59
Unisys-patented Lempel Ziv Welch (LZW) gamma, 59-61
algorithm, 222 HTML, 342
source code, 42-46, 183-185
units
web pages, 51
length, 282-283
websites, 393-401
percentage, 283-284
Virtual Machine, 344-346
UNIX, 180
Virtual Reality Modeling Language (VRML),
updating sites, phases of web projects, 167-168
350-351
upgrades, 18-month pregnancy, 31
vischeck website, 217
URL (Uniform Resource Locator), 6, 180
Visibone, 56
Usability Commandments, 19 executing, 219-221
Usability People, 16 website, 221
User Agent, 253. See also browsers visually appealing, role of web designers,
User Datagram Protocol (UDP), 125 144-145
users Volumeone, 399
control, 36 VRML (Virtual Reality Modeling Language),
of the Web, 8 350-351
utilities
BBEdit, 242
DeBabelizer, 238 W
dicing, 248-250
W3C (World Wide Web Consortium), 24,
HTML, 190-192
117, 256
inline styles, 270
Photoshop, 209-211 Waferbaby website, 336
slicing, 248-250 WaSP (Web Standards Project), 106
Web, 6-7, 115
accessibility, 65, 67
V agnosticism, 23-28
artists, 17
validating HTML, 188-189
communities, 35
VBScript, 27 continuing education, 387-388
vertical type, inserting, 240-243 conventions, 20
video, 38 design. See design
bandwidth, 41 designers, 135
pop-up windows (JavaScript), 307-310 developers, 129-130
quality issues, 38-40 device-independence, 29
18 0732 index 4/24/01 1:04 PM Page 424
424 Web
expanding technology, 7-9 CSS, 257

formats, 210-211 compatibility, 261-262
history of, 14-19, 111-122 content/style, 258, 261
human activity, 35 design methods, 258-259
Liquid Design, 51-55 modification benefits, 259-260
multimedia, 37-40 sizing fonts, 276-284
opportunities, 401-402 strategies, 274-275
practitioners, 131-132 style sheets, 263-270
producers, 131-132 troubleshooting, 271-273
project phases, 151 cutting/pasting, 339-340
rich media, 327-329 elements, 271-272
server-side technologies, 330-331 history of, 14-19
advantages of Java, 349-350 HTML, 175-176
applying SSIs, 341-343 browser compatibility, 204
creating web pages, 331-334 closing tags, 177
cutting/pasting, 339-340 code conventions, 176-177
Java, 343-344 comments, 200-201
Java Virtual Machine, 344-346 constructing tags, 178
middleware applications, 334-335 formatting, 179-181
troubleshooting Java, 347-348 frames, 194-195
standards, 25, 29-30 links, 185-187
statistics of users, 8 META tags, 197-200
strength of, 32 plug-ins, 193-195
technicians, 131-132 sensible markup, 189
typography, 62-64 tables, 193-195
viewer control, 36 tools, 190-192
Web Content Accessibility Guidelines, 66 tutorials, 181-182
URLs, 180-181
Web Design Group’s “Web Authoring FAQ,” 182
validating, 188
communicating through design, 137-138
viewing source code, 183-185
role of, 136-146
WYSIWYG applications, 202-204
B2B, 139-140
inline styles, 269-270
B2C, 140
Java applets, 346
communication, 140-141
JavaScript, 285-286
look and feel of web sites, 138-139
applying, 288-290
medium restrictions, 142-145
Web Developer’s Virtual Library, 188 default status, 299
Web Navigation: Designing the User event handlers, 295-298
Experience, 82 executing, 299-300
Web Page Design for Designers, 393 global documents, 321-324
history of, 287
web pages, 128, 350 image rollovers, 302-306
colors links, 300-301
16-bit, 57-59 platform/browser detection code, 316-320
gamma, 59-61 pop-up windows, 307-310
viewing, 55-57 pull-down menus, 310-312
creating, 20-23 resources, 291-293, 324
18 0732 index 4/24/01 1:04 PM Page 425
websites 425
multimedia, 352 Dave Raggett’s “Getting Started With

SMIL, 352-354 HTML,” 182
SVG, 354-358 David Siegel, 221
plug-ins, 358-362 Digital Web Magazine, 83
Beatnik, 368 disappearing websites, 6
developing content, 376-381 Dreamless, 36
QuickTime, 364-367 drymartini.com, 89
RealPlayer, 362-364 dynamic, 329-331
Shockwave/Flash, 369-376 failure of dot.coms, 80
troubleshooting, 381-385 Flution 1.5, 354
WMP, 367 formatting, 387-388
printing, 88 Fray, 36
resolution, 278-279 Furbo Filters Webmaster, 60
server-side technologies GammaToggle FKEY, 60
ASP/PHP, 335-339 Gap, 76
creating, 331-334 Glassdog, 87
middleware applications, 334-335 Glish, 86
source code, 42-46 Google, 332
SSI, 341-343 Happy Cog, 315
style sheets, 268-269 harrumph.com, 92
viewing, 51 hesketh.net, 100
VRML, 350-351 HTML 4.01 Specification, 183
Web Phones, 29 Ian S. Graham’s “Introduction to HTML,” 181
IBM, 142
Web Review, 394
Icon Factory, 78
websites, 6, 124, 128 Jay Boersma’s Web Work, 181
A List Apart, 182 JavaScript resources, 292
Amazon.com, 77 jazzradio.net, 102
Ask Doctor Web, 181 Lance Arthur’s Design-o-Rama, 182
Babelfish, 35 Lynda Weinman, 221
Barney’s, 106-107 Marc Klein’s Creative Republic, 195
Bloomingdales.com, 101 Mary Quant, 74
Boxtop Software, 229 navigating, 19
Builder.com, 182 overstock.com, 79
caffemocha.com, 103 Photomontage, 17
case studies Presstube, 17
Metafilter, 337-338 Project Cool’s Gettingstarted.net, 181
Waferbaby, 336 publishing, 205-206
clarity, 87 RealSlideshow, 353
brevity, 90-91 Redcricket, 36
icons, 88, 90 resources, 389
structural labels, 90 A List Apart, 390
text, 91-94 Astounding Websites, 390
CNET.com, 77 Babble List, The, 390
Color Blindness Simulator, 217 Dreamless, 391
Computer Currents High-Tech Dictionary, 129 Evolt, 391
creating effective sites, 71-72
18 0732 index 4/24/01 1:04 PM Page 426
426 websites
Metafilter, 391 white space, 91-93

online classes, 392 window.location (JavaScript), 315
Redcricket, 392
window.status (JavaScript), 296
viewing for ideas, 393-401
Webdesign-1, 392 windows, pop-up (JavaScript), 307-310
Ron Woodall’s HTML Compendium, 183 Windows, viewing source code, 185
Sapient.com, 94 Windows Media File (WMF), 367
Section 508 of the Workforce Investment
Act, 66 Windows Media Player (WMP), 367
semantic, 251-252 WinFTP, 205
Transmit, 205 Wired Magazine, 117
useit.com, 16
WMF (Windows Media File), 367
vischeck, 217
Visibone, 221 WMP (Windows Media Player), 367
W3C (World Wide Web Consortium), 24 Word.com, 119
Web Design Group’s “Web Authoring FAQ,” 182 workarounds, 31
Web Developer’s Virtual Library, 188
World Wide Web Consortium (w3c), 24, 117, 256
Web Standards Project, The, 25
Webmonkey, 182 World Wide Web. See Web; websites
webstandards.org, 106 WYSIWYG (What You See Is What You Get), 28,
white space, 91-93 202-204
Web Standards Project, 25, 394
Web Techniques, 394
X-Z
Web Wonk, 117
web-safe color palettes XHTML (Extensible Hypertext Markup Language),
fonts, 245 29, 128, 177
Netscape Color Cube, 215-219 XML (Extensible Markup Language),
Webdesign-1, 392 125-127, 179
Webmonkey, 131-132, 182, 394 Yahoo!, 117
webstandards.org, 106 Yang, Jerry, 117
Webtype, 394 Yenz: The Secret Garden of Mutabor, 400
well-formedness, 179
zooming text, 64
What You See Is What You Get (WYSIWYG), 28,
202-204 Zope, 167
ZX26, 400
19 0732 BM 4/24/01 11:27 AM Page 427
N E W R I D E R S
G R A P H I C S
C L A S S I C S
Designing Web Usability

Jakob Nielsen
ISBN: 156205810X
$45.00
The Authors. The Content. The Timeliness.
Flash Web Design: the v5 remix
Hillman Curtis
What it takes to be a classic.
ISBN: 0735710988
$45.00
Photoshop 6 Web Magic
Jeff Foster
ISBN: 0735710368
$45.00
<designing web graphics.3>
Lynda Weinman
ISBN: 1562059491
$55.00
The Art & Science of Web Design
Jeffrey Veen
ISBN: 0789723700
$45.00
<creative html design.2>
Lynda Weinman and
William Weinman
ISBN: 0735709726
$39.99
Don’t Make Me Think!
Steve Krug
ISBN: 0789723107
$35.00 W W W . N E W R I D E R S . C O M
19 0732 BM 4/24/01 11:27 AM Page 428
www.informit.com
Solutions from experts you know and trust.
New Riders has partnered ■ Master the skills you

OPERATING SYSTEMS need, when you need them.
with InformIT.com to bring
WEB DEVELOPMENT
■ Call on resources from
technical information to your some of the best minds
in the industry.
PROGRAMMING desktop. Drawing on New Riders
■ Get answers when you need
NETWORKING authors and reviewers to provide
them, using InformIT’s com-
additional information on prehensive library or live
CERTIFICATION experts online.
topics you’re interested in,
AND MORE… ■ Go above and beyond what
InformIT.com has free, you find in New Riders
books, extending your
Expert Access. in-depth information you won’t
knowledge.
Free Content. find anywhere else.
As an InformIT partner, New
Riders has shared the wisdom and
knowledge of our authors with
you online. Visit InformIT.com
to see what you’re missing.
www.informit.com ■ www.newriders.com

Taking Your Talent To The Web

Uploaded by

Copyright:

Available Formats

Taking Your Talent To The Web

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Taking Your Talent To The Web

Uploaded by

Copyright:

Available Formats

What is the purpose of the book?

What is the purpose of the book?

What topics does the book cover?

What topics does the book cover?

00 0732 FM 4/24/01 1:38 PM Page i

201 West 103rd Street, Indianapolis, Indiana 46290

Taking Your Talent to the Web:

Trademarks Interior Designer

Part I WHY: Understanding the Web

Chapter 1 Splash Screen 5

Chapter 2 Designing for the Medium 13

Chapter 3 Where Am I? Navigation & Interface 69

Part II WHO: People, Parts, and Processes

Chapter 4 How This Web Thing Got Started 111

Chapter 5 The Obligatory Glossary 123

Chapter 6 What Is a Web Designer, Anyway? 135

Chapter 7 Riding the Project Life Cycle 147

Part III HOW: Talent Applied (Tools & Techniques)

Chapter 8 HTML, the Building Blocks of Life Itself 175

Chapter 9 Visual Tools 209

Chapter 10 Style Sheets for Designers 253

Chapter 11 The Joy of JavaScript 285

Chapter 12 Beyond Text/Pictures 327

Chapter 13 Never Can Say Goodbye 387

Part I WHY: Understanding the Web 3

2 Designing for the Medium 13

It’s Still the Bandwidth, Stupid 45

3 Where Am I? Navigation & Interface 69

Clarity Begins at Home (Page) 87

Part II WHO: People, Parts, and Processes 109

4 How This Web Thing Got Started 111

5 The Obligatory Glossary 123

6 What Is a Web Designer, Anyway? 135

7 Riding the Project Life Cycle 147

Part III HOW: Talent Applied (Tools & Techniques) 173

Doin’ it in Netscape 184

9 Visual Tools 209

10 Style Sheets for Designers 253

Trouble in Paradise: CSS Compatibility Issues 271

11 The Joy of JavaScript 285

Watching the Detection 316

12 Beyond Text/Pictures 327

The Impossible Lightness of Plug-ins 361

13 Never Can Say Goodbye 387

About the Author

About the Technical Editor

Web vs. Print: A Note About URLs

Now go forth, design, and conquer.

He said, “I want to buy you lunch.”

When he told us his company’s vision, we liked that even more.

It was a simple solution to a complex problem. On one side, thousands of

The Populi Curriculum in Web Communication Design, created in coopera-

You know what to do.

WHY: Understanding the Web

2 Designing for the Medium 13

3 Where Am I? Navigation & Interface 69

WHAT DO DESIGNERS DO? Designers organize information, shape identities,

Whether you design websites full-time or just occasionally, you’ll be help-

Before you do anything drastic, such as buying “web software,” changing