Turbo Tutor Version 1.0 Jan85

The TURBO
Pascal Tutor·
A Self-Study Guide to TURBO Pascal
Copyright @1984
Copyright @1985
BORLAND INTERNATIONAL, INC.
4585 Scotts Valley Drive
Scotts Valley, CA 95066
U.S.A.
Borland's No-Nonsense License Statement!
This software is protected by both United States Copyright Law and

International Treaty provisions. Therefore you must treat this software just
like a book with the following single exception. Borland International
authorizes you to make archival copies of the software for the sole purpose
of backing-up your software and protecting your investment from loss.
By saying, "just like a book", Borland means for example that this software may be
used by any number of people and may be freely moved from one computer location to
another so long as there is No Possibility of it being used at one location while it's
being used at another. Just like a book that can't be read by two different people in
two different places at the same time, neither can the software be used by two different
people in two different places at the same time. (Unless, of course, Borland's Copyright
has been violated.)
WARRANTY
With respect to the physical diskette and physical documentation enclosed
herein, BORLAND INTERNATIONAL, INC., ("BORLAND") warrants the same
to be free of defects in materials and workmanship for a period of 30 days from
the date of purchase. In the event of notification within the warranty period of
defects in material or workmanship, BORLAND will replace the defective
diskette or documentation. The remedy for breach of this warranty shall be
limited to replacement and· shall not encompass any other damages, including
but not limited to loss of profit, special, incidental, consequential, or other similar
claims.
BORLAND INTERNATIONAL, INC., SPECIFICALLY DISCLAIMS ALL OTHER

WARRANTIES, EXPRESSED OR IMPLIED, INCLUDING BUT NOT LIMITED
TO, IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A
PARTICULAR PURPOSE WITH RESPECT TO DEFECTS IN THE DISKETTE
AND DOCUMENTAITON, AND THE PROGRAM LICENSE GRANTED HEREIN.
IN PARTICULAR, AND WITHOUT LlMIMTING OPERATION OF THE
PROGRAM LICENSE WITH RESPECT TO ANY PARTICULAR APPLICATION,
USE, OR PURPOSE. IN NO EVENT SHALL BORLAND BE LIABLE FOR ANY
LOSS OF PROFIT OR ANY OTHER COMMERCIAL DAMAGE, INCLUDING
BUT NOT LIMITED TO SPECIAL, INCIDENTAL, CONSEQUENTIAL OR OTHER
DAMAGES.
GOVERNING LAW
This Statement shall be construed, interpreted and governed by the laws of the
state of California.
PREFACE
Congratulations! Since you have this book, you have undoubtedly
joined the ranks of the 200,000+ owners and users of TURBO
Pascal. And you've probably bought this book to help you better
learn how to use TURBO Pascal. That's what this book is for, and
we don't think that you'll be disappointed.
If you're a novice computer user, or novice programmer, then
you'll probably want to begin with Part I, TURBO Pascal for the
Absolute Novice. This will help you to get your feet wet, showing
you just how to get started with TURBO Pascal. If you're a more
experienced programmer, you may want to skim through these
chapters, especially Chapter 6, Getting Started with TURBO
Pascal.
If you know yourway around a computer pretty well, then you're
probably ready to dive into Part II, A Programmer's Guide to
TURBO Pascal. This starts right at rock bottom and builds swiftly,
taking you step by step through all the different aspects of Pascal
in general, and TURBO Pascal in particular. By the time you get to
the end, you should have a solid foundation in Pascal and be able
to glean any additional needed information from the TURBO
Pascal Reference Manual and other books.
Once you feel comfortable with TURBO Pascal, you might take a
look at Part III, Advanced Topics in TURBO Pascal. This section
contains listings of working programs, showing you how to do
things like read the directory off a disk or communicate through a
serial port. The listings also demonstrate different programming
techniques; they're worth studying for that, if for no other reason.
Of course, this book is more than just a book: it comes with a disk
as well. The disk is filled with running programs and tutorial
information, giving you a ready-made library of routines to copy
into your own programs. This is both time-saving and edu-
cational, especially as you adapt these routines to suit your
needs.
iii
This book, of course, can't replace the TURBO Pascal Reference
Manual. Rather, this book's goal is to help you grasp the basic
principles underlying Pascal and its various aspects; the Re-
ference Manual can then help pOint out the exact definitions of
the TURBO Pascal implementation.
Hope you enjoy your exploration. Good luck and have fun!
- Borland International Inc.
Acknowledgements
The following products mentioned in this manual are manufactured and/or sold by the following
companies:
The IBM Personal Computer (PC) is a product of IBM. IBM® is a registered trademark of International
Business Machines Corporation.
MicroPro® and WordStar® are registered trademarks of MicroPro International Corporation.
CP/M® is a registered trademark and CP/M-80™ and CP/M-86™ are trademarks of Digital Research. Inc.
Microsoft® is a registered trademark and MS™ is a trademark of Microsoft Corporation.
Tupperware® is a registered trademark of Dart Industries Inc.
TURBO™ Pascal and SideKick™ are trademarks of Borland International. Inc.
Third Edition. January. 1985

Printed in the United States of America
987654321
iv
ABOUT THE AUTHOR
Frank Borland is more mystique than mystic, as elusive as the
Trinity Alps big foot, as shy as the famous Loch Ness monster.
Even at Borland International, his namesake, few people have
ever seen him. The old-timers recognize him for his remarkable
algorithms, still the fastest in the west. Borland lives deep in the
Santa Cruz mountains with his transportable computer, his
burro, and his dogs. In the early days, he didn't have a permanent
homestead, but lived in a couple of camps deep in the redwood
groves. Now, Frank has settled down a little, bought a cabin, and
is raising a family, thanks to the success of his programming.
These days he is seen even less around town, but still can
occasionally be reached by modem.
If you are a Compuserve user, you are closer to Frank Borland
than you realize. He is writing either a gothic novel or an epic
poem-he hasn't decided which-entirely in bulletin board
messages left on different SIGS (Special Interest Groups). But he
never uses his real name, and he switches names often, so his
writing is hard to follow. Look for messages in cadence, or
rhymes. (You can find information on Borland products and a
Borland SIG by typing GO BOR from any Compuserve prompt.)
Frank is a warm-hearted person. He wrote Sidekick (one of his
latest programming efforts) for humanitarian reasons. Carrying
notepads, calculator, and calendar from camp to camp was
beginning to stunt the growth of his burro, Lotus, so he wrote
Sidekick to make all that unnecessary. He left a note in our
mailbox, saying he'd saved Lotus' development.
He rarely talks about his background, or why he chose to
abandon normal life and take to the mountains. Some say it had
do to with changing the whole motherboard on a PC, just to
replace a single chip. Others blame the high price of micro-
computer software. We don't really know. Do you?
v
TABLE OF CONTENTS
Preface 0 0 0 0 000 0 0 0 0 0 0 0 0 0 0 00000 0 0 00000000 0 0 0 0 0 0 000 iii
About the Author 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 v
PART I TURBO PASCAL FOR THE ABSOLUTE NOVICE

1. Introduction to TURBO Pascal 0 0 0 • 0 0 0 • 0 •• 0 0 0 1-1
2. A Short History of Pascal 0 •• 0000.000.00000.0 2-1
3. Writing a Simple Program. 0 • 0 0 0 0 0 • 0 0 0 0 0 • 0 0 •• 3-1
4. Programming is More Than
Writing a Program 0 0 0 • 0 0 0 0 0 0 0 0 • 0 0 0 0 0 0 0 • 0 0 0 0 4-1
4.1 Data 00000000000.000.00 •••• 00 •• 0000.. 4-1
4.2 Variables 0 • 0 •• 0 0 • 0 0 0 0 0 0 0 0 0 • 0 0 0 • 0 • 0 0 •• 4-4
4.3 Statements. 0 • 0 0 0 0 0 0 0 0 • 0 0 0 0 0 0 0 0 0 0 0 0 0 0 4-5
5. Developing a Program 0 0 0 0 0 •• 0 0 0 0 0 0 0 0 •• 0 • 0 0 0 5-1
5.1 Pseudocode 00.0.0.000000 •• 0 • 0 0 0 • 0 0 •• 5-4
6. Getting Started with TURBO Pascal 0.000000. 6-1
6.1 Turbo Pascal ... 0 0 0 0 0 • 0 0 0 • 0 0 00 •• 0 0 • 0 0 6-2
6.2 The Main Menu 0 • 0 0 0 • 0 0 •• 0 0 0 0 • 0 0 0 0 0 •• 6-3
6.3 Conclusion 0 0.0 •• 00.00.000000000000.. 6-16
PART II A PROGRAMMER'S GUIDE TO TURBO PASCAL

7. The Basics of Pascal. . 0 0 • 0 0 0 0 0 0 0 0 .0 0 • 0 0 .0.0. 7-1
7.1 A Quick Example 00000000000000.00.00 7-1
7.2 Some Pascal Terms. 0 0 0 • 0 0 0 0 000.00.0. 7-5
7.2.1 Characters o. 0 • 0 •• 0 0 • 0 0 0 • 0 0 0 0 0 •• 7-6
7.2.2 Identifiers 000 0 0 • 0 0 0 0 0 0 0 • 0 0 0 0 0 0 0 0 7-8
7.2.3 Reserved Words 0 0 0 0 0 • 0 0 0 0 0 • 0 • 0 0 7-8
7.2.4 Symbols .. 0 0 0 0 0 0 0 .0000000000.00 7-9
7.2.5 Constants 00000000 •• 0 0 00.00 •• 0 o. 7-9
7.2.6 Variables. 0 0 • 0 0 • 0 • 0 •••• 0 • 0 • 0 0 0 0 0 7 -10
7.2.7 Expressions 0000.0.0000.0.0 ••• 00 7-11
7.3 Review. 0 0 0 •• 00000000000.000.00 •• 0.00 7-12
8. Program Structure 0 • 0 0 0 0 0 0 0 0 0 0 0 •• 0 0 0 • 0 • 0 0 • 0 0 8-1
8.1 The Declaration Section ... 0 • 0 • 0 0 0 • 0 0 0 8-2
8.1.1 The PROG RAM Statement 0 •• 0 •• 8-2
8.1.2 Declaration Statements .. 0 •• 0 0 0 0 8-2
8.2 The Program Body 0 00. 0 ••• 0 0 0 • 0 0 0 • 0 0 0 8-4
8.3 Comments 0 0 • 0 0 0 0 • 0 0 •••• 0 ••• 0 • 0 0 • 0 0 0 8-8
vii
8.4 Advanced Program Structure............ 8-12
8.4.1 Typed Constants .................. 8-14
• 8.4.2 Subprograms...................... 8-15
8.4.3 Block Statements . . . . . . . . . . . . . . . . . . 8-17
9. Predefined Data Types.......................... 9-1
9.1 Variables. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9-1
9.2 Integer. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9-3
9.1.2 Integers as Unsigned Values. . . . . . . 9-7
9.3 Byte. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9-9
9.4 Real.................................. 9-10
9.5 Char................................ 9-15
9.6 Boolean.. . . . . . . . . . . . . . . . . . . . . . . . . . .. 9-17
10. Control Structures. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10-1
10.1 Statements. . . . . . . . . . . . . . . . . . . . . . . . . . . . 10-1
10.2 Boolean Expressions. . . . . . . . . . . . . . . . . . . 10-3
10.3 Conditional Execution ................. 10-4
10.4 ·Iteration............................... 10-6
10.4.1 The FOR ... DO Loop............ 10-7
10.4.2 The WH I LE ... DO Loop ......... 10-9
10.4.3 The REPEAT ... UNTIL Loop..... 10-10
10.5 The Case Statement ................... 10-11
10.6 Sample Program....................... 10-14
10.7 Conclusion............................ 10-14
11. Procedures and Functions. . . . . . . . . . . . . . . . . . . . . . 11-1
11.1 Subprograms.......................... 11-1
11.2 Scope................................. 11-3
11.3 Parameters. . . . . . . . . . . . . . . . . . . . . . . . . . . . 11-6
11.4 Functions. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11-8
11.5 Recursive Subprograms. . . . . . . . . . . . . . . . 11-11
11.6 Forward Declarations .................. 11-12
11.7 External Subprograms. . . .. . . . .. . .. .... 11-13
12. Declared Scalar Types. . . . . . . . . . . . . . . . . . . . . . . . . . 12-1
12.1 Subranges............................. 12-12
12.2 Direct Declarations .................... 12-13
13. Arrays......................................... 13-1
13.1 Packed Arrays . . . . . . . . . . . . . . . . . . . . . . . . . 13-5
13.2 Array Initialization. . . . .. .. . . . . .. . . .... . . 13-5
13.3 Order of Elements..................... 13-6
viii
14. Strings........................................ 14-1
14.1 String Comparisons .................. 14-5
14.2 String Functions and Procedures..... 14-7
14.3 Numeric Conversions................. 14-13
14.4 Strings as Parameters................. 14-14
14.5 Stri ngs, Characters, and Arrays. . . . . . . 14-16
15. Records. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15-1
15.1 The "WITH" Statement ............... 15-4
15.2 Variant Records. . . . . . . . . . . . . . . . . . . . . . . 15-8
16. Sets. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16-1
16.1 Set Comparisons..................... 16-3
16.2 Set Operations. . . . . . . . . . . . . . . . . . . . . . . . 16-4
17. Pointers and Dynamic Allocation............. 17-1
17.1 Pointers ............................... 17-3
17.2 The Heap............................. 17-4
17.3 Linked Lists. . . ... . . . . .. .. . . .. . .. . .. . .. 17-6
17.4 Deallocation and
Memory Management. . . . . . . . . . . . . . . . . 17-9
18. Files. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18-1
18.1 Random Access of Files. . . .. . . . .. . . .. 18-5
18.2 Text Files. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18-7
18.3 The EOLN Function .................. 18-9
18.4 Formatted Output..................... 18-10
18.5 Fi lenames . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18-13
18.6 Untyped Files......................... 18-14
18.7 Device I/O ............................ 18-16
18.8 IIReal-Time" Keyboard Input. . . . . . . . . . 18-19
18.9 I/O Error Handling. . . . . . . . . . . . . . . . . . . . 18-20
PART III ADVANCED TOPICS IN TURBO PASCAL

19. Useful TURBO Pascal Routines............... 19-1
19.1 Function Key Detection
(FUNCKEYS.PAS) ..................... 19-1
19.2 Buffered Input (TYPEHED.PAS) ........ 19-2
19.3 I/O Error Checking (IOERROR.PAS) ... 19-3
ix
20. MS-DOS Routines . ............................. 20-1
20.1 Randomize (RANDOM.PAS) .......... 20-2
20.2 Read Directory (DIRECTRY:PAS) ..... 20-4
20.3 Read Directory II (ODL.PAS) .......... 20-6
20.4 Disk Status (DISKSTUS.PAS) ......... 20-11
20.5 DOS Version Number
(VERSION.PAS) ...................... 20-13
20.6 Direct Video Output
(ISMINTIO.PAS) ...................... 20-13
20.7 Direct Memory Output
(MEMSCREN.PAS) ................... 20-15
20.8 Reading the Command Line
(CMDLlNE.PAS) ...................... 20-17
20.9 Seial Port Library
(COMLlS.PAS) ....................... 20-18
20.10 Microsoft Mouse Interface
(TSOMOUSE.PAS) ................... ·20-23
20.11 FiliChar Demo (FILLCHAR.PAS) ...... 20-25
21. CP/M Routines ................................. 21-1
21.1 Read Directory ....................... 21-1
21.2 System Status ........................ 21-2
22. Assembly Language Routines .................. 22-1
22.1 In-Line Code (INLlNE.PAS) ........... 22-2
22.2 Assembly Language Routines
(PASS FUNC.PAS, PASS.ASM) ....... 22-3·
22.3 Conclusion ........................... 22-6
23. Subject - Index
23.1 ......................................... 23-1
x
PART I
TURBO PASCAL FOR
THE ABSOLUTE NOVICE
Part 1 Chapter 1 Introduction
1. INTRODUCTION
I'd like to begin by thanking you for buying TURBO Pascal. Like
me, you must be fed up with the high price of computer software,
and are ready to get good value for (very little of) your money.
Don't be fooled-my software is as good as, if not better than, the
high-priced programs on the market, many of which cost 10
times more than mine. As you can imagine, I am quite proud of
my programs, especially TURBO Pascal.
But before I forget my manners, let me introduce myself. My
name is Frank Borland, and the company which puts out TURBO
Pascal was named after me.
I wrote this book so that you could learn to program in TURBO
Pascal. You and I will look at the Pascal language (that is to say,
standard Pascal), and then I will teach you how to write good
programs in TURBO Pascal.
I live up in the mountains with my family, my dogs, and my burro,
Lotus. It is peaceful here. Just the kind of place where a person
can sit around and think and put things into proper perspective.
Once in a while, when I really want to think through a problem, I
take Lotus and hike a few miles from home. There I will camp out
for a few days until the quiet and isolation have allowed me to
solve my problem. Then I return home to my computer and easily
bang out the code that was such a problem only a few days
before.
Some people think I am an elusive person, with an aura of
mistique surrounding me. Actually, I am just a simple man with
simple needs. I keep to myself for the most part, not because I am
a loner, but because I got tired of city life, and besides, this is the
only way I can write enough good programs to keep the folks at
my company happy.
It was while camping under the stars that I got the idea to write
TURBO Pascal. The thought occurred to me that people like you
wanted to write computer programs with a good programming
1-1
Chapter 1 Introduction Part 1
language, but probably didn't want to pay the high prices

companies are charging for their languages. I couldn't think of a
single reason that you could not have a complete, high-quality,
honest-to-goodness programming language for the price of a
pair of shoes or a decent meal. So, I set out to write TURBO
Pascal. I think it turned out very well, and I use it myself for most
of my daily programming chores.
I don't know how much you know about computers, so I'll just
assume that you are like I was when I got my first computer. I
couldn't make it do anything. (Of course, computers have
become easier to use over the years.) I had to rely on terrible
manuals with missing information, a bunch of individual pieces
that were not designed to work together, and my own patience.
Fortunately, I had a few friends who had already been through
the initial frustrations, and who were willing to help get me going.
It wasn't long before I could do many things with my computer,
and was actually looking for more things to do.
Now that I have been through all of that, I am ready to help other
people like you to get going with such things as TURBO Pascal.
Sometimes all it takes is a certain bit of information presented in
the right way to make you see the light. Other times it takes pages
of explanation and examples. This book is an attempt to bridge
the gap between your knowledge of computers and whatever you
need to know to write good programs.
This book starts out slowly, and then builds upon what you have
learned as you progress. Before I teach you how to program in
Pascal, I am first going to give you a little background on
computers and computer programming. I'll also teach you some
habits which I have found keep things simple and unconfused.
Next we will look at each of the parts of Pascal and why they are
there. We will look at different kinds of numbers and what they do.
We will even learn about a logical algebra called Boolean.
From tim~, to time we will be looking at the TURBO Pascal
Reference Manual which came with yourTURBO Pascal. While it
may be a little confusing for you at this time, it will prove to be a
1-2
Part 1 Chapter 1 Introduction
real treasure chest of information when you know more about

Pascal. It will also serve you well, because you will outgrow this
tutorial pretty quickly.
Besides the Reference Manual, there are a few other things you
will need. First, is your computer-of course!
Since TURBO Pascal is available for so many different types of

computers, I will have to keep this tutorial as general as possible.
In other words, I will not be able to give you a lot of specific
pointers about your particular computer.
For that reason, you should take a little time to get familiar with
how your computer works, if you haven't already. In particular,
you should know how to start a program going (we call this
running a program), save a file, and make backup disks. You
should also know where your CONTROL, ESCAPE, and
Programmable Function keys are (if you have any).
It will also help you in your learning to have a quiet and
comfortable spot to place your computer. There should be room
around it for you to place this book and your Reference Manual,
and enough light for you to read them easily, as well. Learning
anything new is hard enough; you should stack as many things in
your favor as possible.
Something else you will find very handy is a printer. It is almost·
impossible to program effectively without one. The reason is this:
Although TURBO Pascal will find most of your syntactical errors

for you-mistakes where you leave out a punctuation mark, or
call a variable by a different name-it won't find another kind of
error. Every now and then, you will find yourself stumped with a
logical bug-something in your program which won't let it work
the way you think it should. Well, a printed copy of what you have
just done can be compared to a printed copy of your earlier work,
or traced by hand to see just what you have told the computer to
do. Sometimes your mistake will just jump out at you, if you can
see your program on a printed page.
1-3
Chapter 1 Introduction Part 1
Another handy thing about printers and Pascal has to do with the
very structure of Pascal. You will find that big programs written in
Pascal are made up of little programs written in Pascal. We call
these little programs procedures, and we'll discuss them a little
later On in this book. The result is that these little Pascal programs
(procedures) can be stored on your floppy disk for later use, and
they can be printed out and stored as printed programs which
you could then enter when needed.
Then, when you are working on the solution to a problem that
your computer can help you with, you won't have to reinvent the
wheel. It very often is the case that you have already written
something for another program that will be of use in the program
you are working on now. A quick look through your printout file
will be more reliable than an anxious scram ble tryi ng to remem ber
where you last saw the code that you need. So that's about it for
introductions. Let's get on to something I have always found to be
important to my understanding of a new matter-a bit of history.
. 1-4
Part 1 Chapter 2 Short History
2. A SHORT HISTORY OF PASCAL
I can still remember back in the early days of computers (it wasn't
that long ago) when programs were hand-entered by flipping
switches to toggle the state of the computer's electronics. In fact,
before my time, the first computers were mechanical. To program
by flipping switches, you had to know your particular computer's
machine language and you had to convert everything to binary
(base 2 numbering system) or some other number representation.
And, if that wasn't bad enough, you had to check your completed
program by stepping through it and looking at a series of lights.
UIII Iplll ifill! 11l{

1"11. I II 11/1/1/'/11/11
~
~
Bye
-An Early Programming Attempt-
2-1
Chapter 2 Short History Part 1
You see, computers (that is, digital computers) are devices which
only understand different patters of two different states. One
state is the presence of voltage (often called a "1 "); the other state
is the absence of voltage (often called a "0"). Telling the
computer to do one simple thing like "add 2 + 2" involved
entering a series of patterns of ones and zerQs by flipping a whole
bunch of switches. If the answer came up "5," the poor
programmer would have to look at his mess and try to decide
where the mistake was. The situation tended to make some early
programmers crazy. Boy, those were the days. I am glad they're
behind us.
Later, paper tape came along to ease the frustration. The switches
remained on the computer, but paper tape readers were added as
input devices and paper tape punches were added as output
devices. Little holes were punched in the tape which represented
the ones and zeros (ons and ofts) previously entered by switches.
Punched cards, which were originally designed to automate the
weaving industry, and later used to compile the United States
census, also were used in a similar manner to paper tape.
Entering machine-oriented digits to represent computer instruc-
tions was tedious, time consuming, and error prone. As program-
mers become more frustrated, they began using a shorthand,
English-like method of representing these instructions. They
then converted them to machine "code" by associating computer
instruction values with their shorthand notation.
Finally, someone had the idea to write a program which would
make the computer do the "dirty work" of converting the
programmer's shorthand into machine codes. The result was the
first assembler. This was a huge improvement, and assemblers
are still in wide use today; however, the problem is that both the
assembler and its assembly-language programs must be specific
to a particular computer architecture. You could spend months
writing a program for a particular computer, but to use it on
another type of computer you would have to learn another
assembler language and then spend a lot oftime converting your
original program to a new one for the other machine. This
problem was the reason programming languages were invented.
2-2
A programming language is nothing more than a program

which converts a (supposedly) standard series of instructions
into machine-specific code. Therefore, in theory, if you had two
computers and had the same language for each (the languages
are machine specific), you could write a program in the language
and it would run on both computers. This is called portability, and
is a very important consideration these days.
The United States government Department of Defense (DOD)
decided that it needed a portable programming language. This
language would be common to all the people (and their com-
puters) writing programs for DOD.
The language they chose was the "COmmon Business Oriented
Language", or COBOL for short. They chose COBOL because
someone demonstrated to them that the same program could be
run on two different computers without modifications. Because
the government backed it, COBOL has been, and is still, popular
in certain circles.
Notice, however, I hinted that the ability to run the same program
on two computers was notcompletely accurate. Most languages
have a standard set of instructions which will run on any
computer. In addition, they have extensions or enhancements
which let you easily take advantage of special features of a
particular computer. For example, you may have a computer that
has color, graphics, and sound capabilities and another computer
that has none of these. Usually, you must stay away from these
enhancements if you want to write a portable program.
When scientists learned that computers could be helpful in their
calculations, they invented a language which would translate
formulas. They called it "FORmula TRANslator", or FORTRAN,
for short. It is still in wide use today.
This language works pretty well when you have a scientific
problem to solve; one which usually involves a small amount of
data, but lots of calculations. FORTRAN became such a big deal
because International Business Machines (IBM), a manufacturer
of computers with world-wide clout, adopted it as their "official"
language for their big mainframe computers.
2-3
Chapter 2 Short History Part 1
Both these languages, COBOL and FORTRAN, did what they

were supposed to do pretty well. They were quite big, however.
By, this I mean that they needed a lot of room in a computer's
memory -the part of the cc;>mputer which remembers programs
and data wh~le ,the computer is .on. There were other languages,
of course, but these were the two most well known to Americans.
When microcomputers:came along in the early 1970's, they had
too small a memory capacity to allow these compilers to function.
They were also quite involved to learn.
One of the first "simple" languages was the "Beginner's AII-
purpose Symbolic Instruction Code", or BASIC for short. This is
a language that anyone can learn in a hurry, if they don't have a
very complicated problem to solve. You may already know and
use BASIC. Your version may be an interpreter-type language,
which means that instructions can be executed immediately or
interactively. Another attribute of interpreters is that they are
small and efficientin their use of memory, and are easy to move
from one computer to another. Because the interpreter is
essentially a program which emulates a complier and a computer
while the program is running, its programs tend to run much
slower than equivalent complied code.
Also, a person has to be very goodat remembering what is going
on in all parts of hisprogram. BASIC is a hard language with
which to trace what is happening in a program. It is also hard to
trace exactly where it is happening. You can imagine the
implications of this, if you put yourself in the place of a person
having to fix another's program.
To sum itall up, BASIC is a simple, easy-to-use language that
provides a simple solution to simple problems, but which
provides complex solutions to complex problems, due to its lack
of structure.
The other languages we have discussed are compiler-type
languages, which means you can only run your program after
you have completed writing it and have compiled it into machine
language code. TURBO Pascal is a compiler-type language. One
2-4
of TURBO Pascal's unique characteristics is that it is very small,

fast, and easy-to-use. Effectively, it has the speed advantages of
an interpreter during program development as well as the
execution efficiency of a compiled program.
Pascal was written by a computer genius named Niklaus Wirth in
Zurich, Switzerland, in 1970-1971. He had already'written several
other computer languages such as PL/1 and ALGOL. Dr. Wirth
based some of Pascal's concepts upon the work he had done
before.
The reason he wrote it in the first place was to teach his students
how to program a computer effectively. You see, programming
starts with a definition of a problem, then breaking the problem
down into its smallest parts. The last step is to actually write
commands which the computer will understand, and which will
make the computer work until the problem is solved.
Professor Wirth was concerned because he knew that a student's
first programming,language teaches him habits in the same way
that his first love teaches him about life. It colors his views on a lot
of things. Wirth figured it would be better'to start a student out
with good programming habits.
Pascal is a structured language, which means that it is easier to
write your progarm in modules by following certain, predefined
steps. Certain parts of your program must be placed in certain
locations within the program, and must follow certain conven-
tions. I will teach you more about structured programming as we
journey through the following chapters.
Structured languages are becoming more and more human
oriented and problem oriented. They have the ability to have a very
English-like content. At first it won't be apparent, but as you learn
to read Pascal programs, you will find that the statements read
very much like English sentences. The result is that the problem
and its solution is very easy to see, and is machine independent.
We'll talk more about the technique of programming in the
chapter after next. For now, let's look a little more closely at
Pascal, and especially TURBO Pascal. '
2-5
Part 1 Chapter 3 Simple Program
3. WRITING A SIMPLE PROGRAM
Did you know that computers can show up a basic difference

between people's learning habits? Many people who sit down at a
computer Wllich they have never see before. will start right in
pressing keys. On the other hand, some people will wait for
instructions before touching anything. Even with instructions, it
will take these people quite a while before they are comfortable
with a computer, while the first type of people were never really
uncomfortable in the first place.
Now, the reason I am bringing this matter up here is not to scare
you, or make you feel as though "my goodness, this is too
difficult." It is because I want to encourage you to sit down and
press computer keys, especially if you are the type who is over
cautious about such things. This is the way you will learn about
your computer most quickly. By the way, today's computers are
designed so that with appropriate precautions (such as making
backup copies of all your diskettes), it is highly unlikely that you
will cause any permanent damage by typing on the keyboard.
You see, a computer, unlike your mind, does not think in any real
sense. At its most fundamental level it thinks in terms of "on/off",
or "yes/no", but never "semi-on" or "maybe". It will only follow its
instructions (whether they make sense, or not) until its electrical
power goes off.
3-1
Chapter 3 Simple Program Part 1
DUMB
MACHINE
j
l
aYB
Also 'you should remember that as a human being you have

ultimate control over your computer: you tell it what to do and
when to do it. You do this by pressing keys and giving it
immediate instructions, or by pressing keys for instructions
which the machine will remember and then execute later on. (By
the way, the second example is what we mean when we talk
about writing a computer program).
When we work with cqmputers we have to get into two habits.
One: we a/ways (not sometimes) must be consistent in how we
tell the computer to do something. (In other words, the computer
is very inflexible.) Two: If we give the computer the wrong
information to work with, it won't know the difference, and will try
to proceed as usual, causing erroneous and sometimes interest-
ing results. This is what we will call "Garbage in, Garbage out."
3-2
In other words, with computers, as with no other part of life, it will

always benefit you to take the time to first: "Think It Through"
-to pretty well know what you are going to do before you start,
and second: "Do It Right The First Time" -so you won't have to
do it again.
Well, what do you say we just try a program? Our first program
doesn't do much, and you may not understand everything about
it right away -but then even driving a car is not second nature at
first. If you need to, read the chapter on "Getting Started" in your
Reference Manual to show you how to write and run a TURBO
Pascal program.
By the way, I have written all the special words TURBO Pascal
recognizes as instructions in bold letters. This makes it particu-
larly easy to read a program and make sense out of it. By the way,
these special words are called reserved words.
To get started right away, do the following steps:
1. Start your TURBO Pascal program by following the instruc-
tions in your Reference Manual. Press Y when you are asked
whether you would like messages included.
2. Press Wand answer MYNAME when asked for the name of
your workfile. (Don't forget to press RETURN after you have
typed in the program's name.)
3. Press E to get your TURBO Pascal editor going.
4. Type in the following program, just as it is written here. Don't
forget to put in everything, including all punctuation marks
and spaces. If you make a mistake, you can use your <Del>
key (or whatever key you have defined during TURBO Pascal
terminal installation) to backspace and erase the characters
you've typed.
3-3
program MyName;
const
TotalTimes = 20;
var
Name String[ 25 ] ;
NumberOfI'imes Integer;
begin
Write('What is your name, please: ');
Readln(N arne);
ClrScr;
for NumberOfI'imes := 1 to TotalTimes do
begin
Writeln('Your name is ',Name);
end;
end
5. When you have finished typing in the program, press your
CONTROL key and hold it down while you press K, then let
both keys up and press D.
6. Press S to save the program you just typed in. This way you
will have it later, if you want it.
7. Press R to compile the program (don't worry what that means
for now) and to run it. You will see a message at the bottom of
your screen which you can also ignore for now.
In a moment, the program will run itself.So, just what did this
program do?
If you typed it in correctly, and then typed your name correctly
when it asked you for it, the program told you your name 20 times.
If you did not type it in correctly, or if you did not type your name
correctly, all the program gave you was garbage (if anything at
all). Rememberwhat I said about "Garbage In/Garbage Out?" (If
you did get garbage, you may want to go back and compare what
you typed into your computer with the program here in the book,
and then try again. If you made a mistake and haven't yet learned
the editing commands, it might be easier to retype the entire
program at this time, using a new file name.)
3-4
Now let's look at the different parts of the program and talk for a
moment about each. We will go into more detail about each in a
few pages. For now, just try to get a sense of what is going on.
The program, called "MyName", has three main parts:
First: the place where you tell the compiler the name of the
program.
programMyName;
The compiler uses this information to know where all the

information it needs begins. It is the same as when you want to
give a friend some instructions and you say, "This is where you
start." We call this the Program Header.
Second: the definition statements:
coDSt
TotalTimes = 20;
var
Name string[ 25];
NumberOfTimes Integer;
The compiler uses the definition statements as tools when it

begins to do what you want it to.
const is short for constant, the name of a number which will not
change while the program is running. Since the number stays
constant, we call it a constant-get it?
var is short for variable. Like its name, a variable can contain just
about anything, for instance a name, a number, orsomething you
make up.
3-5
Third: the program statements themselves:

begin
Readln(Narne);
ClrScr;
for NumberOfI'imes := 1 to TotalTimes do
begin
WritelnC'Your name is ',Name);
end;
end.
The program statements tell the compiler specific steps the

computer must take to accomplish its assigned tasks.
An English narrative of this program would go something like
this:
1. Start here
2. Ask "What is your name, please?"
3. Read the answer.
4. Erase the computer's screen.
5. Start a process where you write "Your name is" and the
answer you read before (in step 3) twenty times, then stop the
process.
6. Finish, end, quit, stop, fini.
By the way did you notice that every specific idea, or statement,
ends with a semicolon (;)?This is similarto the way we speak with
each other, in sentences, I mean. We express our ideas, each in
turn, in sentences; Pascal does the same in "statements". We end
our sentences with periods; we end our Pascal statements with
semicolons.
And yes, I see that begin, and const, and var don't have
semicolons. This is an idiosyncrasy of how the compiler operates.
Now let's look at the process of programming, itself, a little more
closely in the next chapter.
3-6
Part 1 Chapter 4 Programming Discussion
4. PROGRAMMING IS MORE THAN

WRITING A PROGRAM
The process of programming a computer is not really about

writing a series of instructions for a machine to follow. Rather, it is
about solving a problem, making life easier, or doing something
you want to do. You accomplish these things through the use of a
tool called a computer-and another tool called your common
sense.
No tutorial would be complete without my giving you some basic
(as in fundamental-not as in BASIC) pointers on how to go
about programming. That's what I will start to do in this chapter.
By the way, when you read this section and then continue to
other sections, I want you to get a sense of what is going on. Don't
try to memorize facts and figures. Once you get a feel for the
basic concepts, you will be able to handle any programming
problem.
-Just Sitting Back Thlnklng-
4-1
Chapter 4 Programming Discussion Part 1
Computers are pretty consistent machines. If they do what you

want them to, they are consistently good. If they don't, they are
consistently bad. The problem arises when we don't realize that
they do exactly what we tell them to- whether or not what we tell
them is what we really want them to do.
Now is that confusing enough for you? Don't worry, it will all be
clear in a little while. What we need to do now is to get
comfortable with some concepts which have to do with computers
and programming.
4.1 DATA
Someti mes it seems that no matter where we go, we are

bombarded with something called data. Data this and data that,
but never anything really specific. People seem to forget that
what we need to get something done, orto make a decision, is to
reduce the data into more manageable pieces. We are going to
define this reduced, meaningful data as processed information.
When we get right down to it, data is the basis from which we
derive information. In other words, we take raw data, and then
manipulate it until we can use it to inform us. Finally, we use this
processed information to make a decision.
If we have done the right things-obtained the appropriate data
and then interpreted it correctly-our decision will work out for
the better. By the way, interpreting data correctly means doing it
with a good deal of horse sense; not with blind trust in the data.
Let's say you want to buy a car and then sell it. Raw data, in this
case, would be the price you could buy it for, and the price you
could get for it. Processed information, however, is the result of
comparing the two. You base your final decision to buy and/or
sell on that information.
Now, the color of the car and its make and model may also be
items of data which need to be considered. The thing is that they
may not really be important (necessary to your decision) in this
case, unless you are interested in such things.
4-2
If it is a simple decision as to whether or not to have a car, your

data can be restricted to simple price considerations. You may,
however, also want to consider make, model, color, gas mileage,
and so forth, depending upon your needs and desires, and the
perceived needs and desires of the buyer.
The bbttom line is this: use common sense to decide what data to
bother to col/ect, what data to use, and what data to trust. Let's go
back to our prices, for a moment. Let's also say that they are all
you choose to consider.
Your two items of raw data (the purchase and selling price) are
nothing more than numbers until you put a $ in front of them. The
$ is what we would call a data label. It, paradoxically, turns the raw
data into processed information-sort of.
A data label helps us use the data in a meaningful way. It does so,
because it takes a number, or bunch of letters, without apparent
intrinsic significance, and jogs our memory.
The data we choose to consider, and which can be manipulated
by a computer, comes in two forms: numbers and letters. A
computer scientist would call the two numeric and character.
Imagine, for a moment, that you see a big plant with green leaves
and shade in front of you. You store this information in your mind
as a picture of what you saw. Later, when someone asks you what
it was that you saw, you would say"a tree." In this way, you would
convey to them the idea of a big plant with green leaves and shade.
This is very similar to how a computer works, inside its "mind."
The difference between you and a computer, is that while you
would never confuse a tree with a cat (because each means
something different to you), numbers and letters mean absolutely
nothing, in themselves, to a computer. It doesn't matter in what
combination the letters or numbers may be.
In other words, since the computer does not care what it takes in,
we can put in almost anything. All we need to do is make the data
going in mean something to us. This way it will also mean
something to us when we get it back out.
4-3
Chapter 4 Programming Discussion Part 1
As a practical matter, we divide the data to be stored in a

computer under two rules:
1. Each character on your keyboard produces a representative
code when pressed. If a character (numbers included) is
simply to be stored as its representative code, we call it
character data. This would include things like names, ad-
dresses, phone numbers, and so forth. (Since we often string
a bunch of characters together into a word, we will refer to a
bunch of characters with a particular meaning, when they are
together, as being a string. If the character makes sense by
itself, we will simply call it a character. Simple, isn't it?)
2. If the data consists of numbers which are to be manipulated,
as in a calculation, and then returned to you, the computer
(complier) coverts the representative codes for the numbers
into their actual values, in whichever number base you select
(decimal, hexadecimal, binary). We call data in this form
numeric data. These are things like dollar amounts, inventory
counts, and so forth. (A date could fall into either category,
depending upon whether or not it is data which requires
manipulation.)
One convenient thing about computers is that because the two
kinds of data fall into clearly defined categories, they can not be
substituted for one another.
Another convenient thing, at least in Pascal, is that sometimes we
have to use data in more than strictly numeric or character form.
For instance, our data could be more than simply a list of days
(Mon, Tue, Wed, Thu, .Fri, Sat), or the alphabet (A through Z).
Pascal allows us to define our data as being almost anything we
want. We will learn more about this later.
4.2 VARIABLES
In orderto store data items in a computer, we need places-similar

to pigeon holes-to put them. We call these places variables, and
this is how they work:
4-4
Let's say we have to store a list of names. The first actual name
(Joe Blow) would be stored in a variable-the pigeon hole-
called NAME[1 ].The second name (Dudley Doright) could go
into a variable named NAME[2], and so on. If we had no data to go
into a variable, it would remain empty.
Variable Contains Character Data
NAME[1] Joe Blow
NAME[2] Dudley Doright
NAME[3]
NAME[4]
You will notice that NAME[1] and NAME[2] aren't worth a thing,
by themselves. All they are good for is holding something else.
(NAME[1] is not Joe Blow, it only holds a string of characters
which means "Joe Blow"). To make things easierfor us, however,
in Pascal we call ourvariables by some name which will give us an
idea of what data they are likely to contain.
4.3 STATEMENTS
When we want to convey to one another one thought, or a

discrete part of one thought, we use a sentence. The same kind of
thing in Pascal is called a statement. English sentences end in
periods; Pascal statements end in semicolons.
For now, it is only important that you understand that when I refer
to a Pascal program statement, I am speaki ng about one clearly
stated instruction (thought) to the computer.
I will give you more definitions later, as we go along in this
tutorial. No need to burden ~/OU with more than is necessary, right
now.
Next we will take a look at the first step in developing a program.
So, if you are ready ...
4-5
Part 1 Chapter 5 Developing a Program
5. DEVELOPING A PROGRAM
By this time, I hope you are getting the impression that a

computer is no more than a tool. It is a singularly impressive and
useful tool, but it is nonetheless a tool in the same way a hammer
or saw is a tool. The only reason to use a tool-or a computer-is
to solve a problem; to make our life easier.
A hammer and saw solve problems for carpenters. After all, it is
easier to saw a board into pieces than it is to rip it apart with your
hands and teeth.
The analogy can be carried further:A carpenter doesn't simply
start in sawing and banging. The first thing he does is to plan and
measure. He knows what he is going to do before he does it. He
also knows which tools are appropriate to his plan.
-Planning, the First Step-
5-1
Chapter 5 Developing a Program Part 1
The best way to go about developing a program is to do so

systematically. Leave the computer alone for awhile, and get out
a pencil and a pad of paper. You need to make a list of things to
do, and then follow that list as you do the actual program entry,
like a pilot following a checklist, or someone walking up a flight of
stairs-one step at a time. It is important to write your plan down
to make it easy to keep track of.
And, perhaps most importantly, you will want to keep in mind
common sense. I n other words, if you have to go back down a few
stairs to pick up something you missed: do it, before you get too
far away.
Here are the steps that folks who program computers recommend
following:
First Step: Identify the Problem to be Solved
In order to solve a problem, you must first have a very good idea
about what the problem is. If your car has a flat tire, the problem is
obvious. So is the solution. But then, a computer isn't much help
in fixing a flat.
On the other hand, if you suddenly are asked to keep track of all
the names and addresses of the families attending your church, a
computer would be ideal. The problem is also reasonably
obvious.
You will have to store data regarding the church members. This
data will involve such things as each member's name, address,
phone number, and maybe birthday.
Part of your identification of the problem should include a
description of why you have to solve it. It should also include an
indication of what results you expect to achieve from your
solution.
You should also take your problem apart, breaking it into its
smallest parts. Then you can attack each part of the problem in its
turn. When each small part of the problem is solved, and you
5-2
Part 1 Chapter 5 Developing a Program
put all of them together, the large problem will be solved. In

computer jargon this is called Top-Down Design and Stepwise
Refinement.
By the way, this method of program design is ideally suited to

TURBO Pascal programming. You will see how when we get into
procedures, a little farther in this book.
One general note:While you are thinking about the problem to be

solved, try to think of a similar problem you have solved before.
This may give you insight into your current problem and suggest
some possible solutions.
Second Step: Decide What You Need to Get Out of Your System
The only reason to store data in your computer is to get it back

out again in the form of information. The only way to know what
data to store is to know what information you will need later.
The best way to go about it is to fully describe each piece of data.

You will want to include the data's source, its form (numeric or
character), and a name for a variable to hold the data. Finally, you
should determine whether or not it makes sense to actually
include each particular data item you have planned for.
Third Step: Decide What Data You Need to Put Into Your System
Sometimes people put deciding what data to put into their system
in front of the second step. When they do, they soon find
themselves "backing down the stairs" to get things right. This is
because getting the information out is basic to knowing what is to
be put in.
As you are deciding what data to put into your system, make sure
that each item is, in fact, available. Then look at it with an idea as
to what mistakes people could make as they enter this data.
For instance, a date entered by an American would normally have
the form: month, day, and then year (MM/OO/YYYY). A European
would more likely put the day before the month as in (OO/MMI
YYYY). Part of good programming is to identify possible errors
before they are made, and to somehow plan around them.
5-3
As another example, let's say that your program is going to ask

the user to answer a question with either a "yes" or "no." Will you
have the user enter "Y" or "N"? What if he instead enters "y" or
"n" or "Yes" or "NO" or "no"? Will your program still work
correctly. Or, what if the user enters "R"? Will you provide an
error message and ask them to enter the answer again?
These are the sorts of things you must think about when
designing a program. Fortunately, TURBO Pascal provides easy
ways to handle these situations, which we will cover when the
time comes.
Fourth Step: Miscellaneous Stuff
This is where you put down reminders to yourself about things

which don't fall neatly into the other three steps. Things like how
many times to do something, how the screen should look, or
anything else. I view this step as making notes on a piece of
scratch paper.
5.1 PSEUDOCODE
As you are writing your plan on paper, you should state each step
in complete sentences. Then, as you write the solution to each
step, each should also be a complete sentence.This collection of
sentences which describe the solution to a problem (in easy-to-
read steps) is called Pseudocode. The term is derived from jargon
for writing a program, which most programmers call coding a
. program, or writing code.
This brings up two more computer terms with which you should
be familiar:SOURCE CODE and OBJECT CODE. Source code
refers to the words you write in the Pascal programming
language using yourTURBO Pascal editor. This is the information
the compiler translates into object code. Object code, in turn, is
the information your computer uses to solve your problem. You
see, you can read source code, while your computer cannot. On
the other hand, your computer can read object code, while you
can't. It is your compiler (the "heart" of TURBO Pascal system)
which does the translation from source code to object code.
5-4
Part 1 Chapter 5 Oeveloping a Program
Do you remember the program we ran back in Chapter 3? Let's

take another quick look at it.
The problem we wanted to solve-one of the world's least
significant-was to see our name a number of times on a
computer screen. Using the method outlined above, we went
about it in the following way:
Problem to be solved: an unreasoning desire to see my name in
lights. Since computer letters are in "Iights" and are cheaper than
a marquee, they will have to do.
Information to get out of the computer: my name, a string of no
more than 25 characters, in a variable called "name".
Data to go into the computer: my name, to go into a stri ng
variable called Name. I get it from my memory, and it is usually
readily available.
Miscellaneous Stuff: I want my name to appear 20 times; I'll use a

constant variable called Tota/Times to hold that number. This
way, if I have to change the number of times sometime later, and
my program grows bigger with Tota/Times appearing in more
than one place, I will only have to change one thing to change
them all.
I will need one more variable called NumberOfTimes to keep
track of the question, "Has the computer written my name 20
times yet?"
I will also want to start with a clean screen, but TURBO Pascal has
already taken care of that for me with a function (about which
we'll talk more later) called C/rScr.
Pseudocode:
1. Start the program

2. Ask "What is your name, please?"
3. Read the answer into a variable called Name and don't let the
answer be longer than 25 characters.
5-5
4. Erase the computer's screen.

5. a. Add 1 to whatever number the integer (which means a
single whole number) variable NumberOfTimes holds.
b. Compare the number now held by NumberOfTimes with
the number held by the constant Tota/Times.
c. If NumberOfTimes is less than or equal to Tota/Times
then write "Your name is" and the string variable Name
and go down to the next line on the screen.
d. If NumberOfTimes is more than Tota/Times, then stop.
To save your having to turn back to where this program first
appeared in this book, here it is again. Compare it with the
pseudocode, and see how close they are: ..
program MyName;
const
TotalTimes = 20;
var
Name : string [ 25 J;
Numbe:r{)1Times : Integer;
begin
Readln(Name);
ClrBcr;
for Numbe:r{)1Times:= 1 to TotalTimes do
begin
Writeln('Your name is ',Name);
end;
end
The only real difference between the two is that the program
source code is written in a way that the compiler will be able to
read it and then translate it. Other than for that accommodation to
the compiler's requirements, it is very close to English.
5-6
Part 1 Chapter 5 Deve/oping a Program
When you get more familiar with TURBO Pascal, and with good
programming habits, you will be thankful for Pascal's similarity to
English. You will be able to take almost any Pascal program,
written by almost anyone, and figure out what they are trying to
do.
Well, that's enough about getting ready to write a program. I
could have continued longer, but we have many more issues to
discuss. Let's get into the next chapter now and get down to the
n itty-g ritty ...
5-7
Part 1 Chapter 6 Getting Started
6. GETTING STARTED
WITH TURBO PASCAL
Sometimes it seems that the hardest part of doing anything is

simply getting started. Then sometimes it seems impossible to
get something finished once you have gotten started.
Now we have come to the time when we are ready to really get
into TURBO Pascal.
The first thing you want to do with TURBO Pascal is to make a
backup copy of diskette you bought. (You did buy it, didn't you?)
Then you will put your store bought, or master, diskette in a safe
place away from kids, stray cats, magnets, and wires with
electricity running through them. Keep in mind that, for now, we
are talking about the TURBO Pascal diskette; not the diskette that
came with this book.
I've made some assumptions here, and I hope that they are
correct. First, I've assumed that you have learned how to do some
of the basic functions with your computer, such as turn it on and
off, bring up its operating system (DOS), usethe keyboard, make
copies of diskettes, and display a directory of files on a diskette. If
my assumption is wrong, then get some help or do some reading
and learn how to do these things before continuing.
When you ask your computer for a directory of programs on the
TURBO Pascal disk, you will notice one named TINST. This is
what we call a general installation program. It is a program which
allows us to use TURBO Pascal on a n umber of different brands
of computers, and yet have it run in the same way on each.
You can skip TINST and go right into writing a TURBO Pascal
program, if you want to. But if you are not using an IBM-PC or
compatible machine and want to use the built-in program text
editor, you will have to run it anyway.
6-1
Chapter 6 Getting Started Part 1
6.1 TURBO PASCAL
To start your adventure in good programming (did he really say

that?) you will want to do the following:
1. Insert a backup copy of your TURBO Pascal program disk in
drive A:. If you want to keep files of programs you write on
another disk, put another blank, formatted diskette in Drive
B:.
2. Type the following (note that the word <Enter> enclosed in
"angle brackets" means press the Enter or Return key on your
keyboard; do not type the word Enter or the brackets themselves):
TURBO <Enter>
3. When the computer asks whether you want the messages included
press:
y
This has to do with mistakes you make when you write your
program and then try to compile it.
By preSSing Y you are telling the computer to tell you what the mistake
was, in English. Pressing N, for"NO", tells the computerthatyou would
rather look up your mistake in the Reference Manual from the mistake's
number (we call this an error code). You can do this because the
computer will tell you the error code number to look up. All in all, it is
simpler, right now, to have the computer tell you.
If you have done the foregoing correctly, you will shortly see what we
call the Main Menu.
6-2
6.2 THE MAIN MENU
When we say menu we usually mean a Iist of items from which we

can pick one and do something with it. I n a restaurant we eat what
we pick. With a computer, we either do something with our
choice, like run a part of a program, or give the computer some
information. With the Main Menu we do the latter. Here is what it
looks like:
Logged drive:
Work file: Mamfile:
Edit Compile Run Save
eXecute Dir Quit compiler Options
Text: 0 bytes
Free:nnmm bytes
>
, Let's take each part of the menu and see what it does. We will do
so in a more or less logical order, but first... do you notice that
some letters are brighter than others? With computers, this
usually means that pressing a key with this letter will have some
effect different from pressing another key. (Notice that there are
no two bright letters the same). That is the case here. When I
mention any option from this menu, it will mean that you should
press a key which corresponds to a bright letter to indicate that
option.
Logged drive:
The logged drive is the disk drive (or hard disk directory) with
which you wish to work. For instance, if you have your TURBO
Pascal master disk in drive A:, and the programs you are writing
stored on a disk in drive B:, then the logged drive should be drive
B:. To make it so, do the following:
6-3
1. Press
L
to choose the Logged drive option on the menu.
2. Press
B
to indicate your choice of drive B:.
3. Press
<Enter>
to tell the computer to accept your instruction.
But how about with a hard disk, or even a floppy disk with
subdirectories? Well, if you're using PC-DOS or MS-DOS,
version 2.0 (or later), you would see the following:
Logged drive:
Active direotory.
This shows the path name of the current default directory. To

change this, you would:
1. Press
A
to indicate the Active Directory option.

2. Type in the name of your disk subdirectory. For instance,
\PASCAL \PROGRAMS, if you haveonedirectoryforTURBO
Pascal and another for the files of programs you write.
6-4
3. Press
<Enter>
to tell the computer to accept your instruction.
Work File:
Main file:
Since these two are so closely related, let's look at them together.
Work file: is the name of the file which is also the program you are
now writing. If you just enter the name of a file, TURBO Pascal
will add the file extension .PAS for you. If you have an extension
in mind which you would rather have than .PAS, just put it in
(TURBO Pascal won't mind).
Main file: has to do with a more advanced programming concept
of INCLUDED FILES. This is farther along in programming than
we will go in this tutorial. Suffice it to say, when you get better
(more knowledgeable) at computer programming, the INCLUDE
FILES facility will be available to you.
If you ever choose to edit, run, or compile a program without
having first indicated a work file, you will be prompted for a work
file name.
Edit
There has to be a way for you to actually write your program in a
form the computer can use. More importantly, there must be a
way for you fix the mistakes you make as you are writing the
program as well as after you have written it and tried it out. A
correct term for this process is "Editing". To accomplish this
process we use an Editor, which is called for our use by pressing
<E>.
6-5
~\n/~·
"/ \f--:-
'BYE.
- TURBO Pascal Edltor-
Note
TURBO Pascal is a compiler-type language in which you
write your programs using an editor, then compile them so
that the computer can understand them. In contrast, most
BASIC programming languages are interpreter-type lan-
guages in which you write your program using the interpreter
and the computer translates it as you go. If you have been
using BASIC, try to forget everything you know about it, it
really won't help you very much with TURBO Pascal.
Writing your program using the TURBO Pascal editor is a lot like
writing a note with a word processing computer program called
WordStar. If you use that program already, then you are familiar
with the various commands which help you accomplish what you
would like to do, and you are all set to go. On the other hand, if
you don't-you're not; but you are still not out of luck.
6-6
In the latter case, you have two choices: either you get familiar
with the commands as they are listed in the TURBO Pascal
Reference Manual, or you "customize" the commands to suit
yourself.
The latter is not a hard thing to do. Just follow the directions in the
Reference Manual Section 1.6.3 Installation of Editing Commands.
By the way, I would make my notes in the Reference Manual in
pencil. Believe me, you will want to keep track of what pressing a
particular key, or key combination, does. Sooner or later you will
change your mind about some command or other. A pencil will
make changing your mind a lot less messy.
When you start to edit a program (by pressing E), the screen will
go blank for a moment while TURBO Pascal is taking care of
some housekeeping. Like anything with TURBO Pascal, the
moment will be a short one because TURBO Pascal does things
so quickly.
What is happening is this: If you have specified the name of a

work file which is already on your disk, TURBO Pascal goes to
the disk and retrieves yourfile. Then it turns on the editor and lists
out your program. Very shortly, it is ready for you to work with
your file. Then you may add stuff, change things around, or do
whatever you need to do. (Note that if you had already edited
another file, you would have been warned that you 'should save it
before editing another.)
If you have specified the name of a file which is not on your disk,
TURBO Pascal will take just a moment to check, and then will
send you into the editor so you can start your programming.
Once you are done editing, you can save your work to disk. (See
Save, below.)
To write a program, once you are "in the editor", is simply a
matter of typing away! Oh yes-to get out of the editor and back
to the Main Menu, press
<Control> KD
6-7
This means, press and hold down the key marked "Control" or
"Ctrl", then press "K", then "0". That's all there is to it.
Compile
The whole object of programming in TURBO Pascal is this:to
write code and compile it. Compiling is the step where you find
out how well you did. This is where you find out if you wrote the
program correctly, and if your program's logic is correct.
B 5
- The TURBO Pascal Compiler-
When you press C, TURBO Pascal takes your edited (source

cOd.e) program and changes it into a form the computer can use.
This form is called object code or compiled code. (You will
remember we spoke about compiled programs in the section on
programming in general).
To go over some jargon again: the compiler takes your source
code, the program you wrote in TURBO Pascal using an editor,
and changes it into object code, the form the computer can use.
6-8
The TURBO Pascal compiler compiles programs, and then puts

them into one of three places: into memory, into a .COM file (or a
.CMD file for CP/M-86), or into a .CHN file. Which of the these
threeto use is up to you and can be changed at any time. You can
see how to implement your decision in the section on "compiler
Options", below. Either way, you compile your program by
pressing C.
Usually you will first do what we call compile into memory. You
do this when you write a program using the editor and then want
to try it out real quick. With the program in memory, you will be
able to run it, find a mistake, and then go back to editing it, if you
want.
If you have made a mistake which is bad enough that TURBO

Pascal can't get around it, the compilation process will stop. Then
TURBO Pascal will switch you back into the Editor, tell you what
it thinks you did wrong (or what you forgot), and show you where
it thinks the problem is.
I
I
I
i
I
I --'-~"'--'-
I, -
!--.. , \
-
COJ1PILD'I.
CArCHE.5
I /c~ :-:.~ £R.R.DP...!
SEND5 IT BliCK
70 EDITOR..-+
-Error Handling-
The thing to do now is to fix your mistake and try again. Believe
me-the first time you write a complete Pascal program, on your
own and then run it will be a high point in your life indeed.
Remember, however, that the compiler will only find a syntax
error. Logical errors should already have been eliminated in your
planning (it says here). Believe me, it is better that they be found
6-9
by you in planning, than by your friends when you show your

program to them. This method also takes a shorter amount of
time, in the long run .
. You can compile your program directly into memory by pressing
C as many times as you want (we'll talk about what that means in
just a moment). One thing: if all you do is'press C, then all you will
be able to check is if the program compiles correctly all the way
through. This is a good idea, as long as there is the possibility of
syntax errors. (Syntax errors are errors such as where you have
left out some necessary punctuation, or made a mistake with the
spelling of a variable name.)
Run
Once your program has compiled all the way through -or if you
are impatient like me, as soon as you are done editing-you will
want to press R which means "Run the program". With this
command, TURBO will see if your program has already been
compiled, or even if you have edited it since you last compiled. If
so, then it will first compile your source code into memory.
Having donethat, it will then immediately execute your program's
instructions.
compiler Option
Now seems like a good time to clarify the difference between
compiling to memory and compiling to a .COM file.
Once you are satisfied with the program, you will want to save it in
a form where you can use it without first having to call up TURBO
Pascal. This is done by compiling your program into a .COM file;
a program file you can run by simply typing its name.
You choose this option by pressing 0 at the Main Menu. With
that, you are presented with another menu which looks like this:
compile - > Memory
Com-file
eRn-file
Find run-time error Quit
6-10
Note that if you are using a CPM-86 version of TURBO Pascal,

the C option on the screen will be to compile a .CMD file. Now,
let's look at each of these options.
Memory
The normal, or default, value is the first one: Memory. In other
words, without any change on your part, TURBO Pascal will
always compile your program into your computer's memory.
Once you have made your choice, pressing Q will send you back
to the Main Menu.
Com-file (or Cmd-file)
If you are at this menu, and you press C, for Com-file (or Cmd-
file), your program will be compiled and placed in a .COM file on
your logged disk drive. If your work file is called "MYNAME", and
you have chosen to compile the program into a .COM file, then
MYNAME will be saved to disk under the name MYNAME.COM.
From then on, you will be able to treat MYNAME.COM (or
. MYNAME.CMD) like any other program file. You will be able to
copy it to another disk, rename it, delete it, or run it, all
independently of TURBO Pascal.
The reason you are able to do this is because your MYNAME
program is not the only thing which went into the MYNAME.COM
file. A part of the TURBO Pascal language, called the library, was
also saved.
A library is a collection of programming routines which are used
by any Pascal program to work with your computer. They are
always there, even if you haven't written them, or ever knew they
were there. The programming routines take care of things like
putting letters on your screen, calling the Disk Operating System
to save information, and so forth. The library will increase the size
of your file by about 10 thousand bytes (10 Kbytes) or so.
When you press C, you will also be prompted for some additional
information. The prompt will look like this:
6-11
minimum cOde segment siZe: XXXX paragraphs (max. YYYV)

minimum Data segment siZe: XXXX paragraphs (max. YYYV)
mInimum free dynamic memory. XXXX paragraphs
mAximum free dynamic memory. XXXX paragraphs
This is an introductory tutorial, and what is going on here is a bit

beyond our current efforts. This information is discussed in the
Reference Manual, if you are interested in finding out more. For
now, you can ignore this stuff.
cHn-file
The Chain File option is selected by pressing H at this menu. A

chain file is similar to a .COM (or .CMD) file, except that it does
not contain anyofthe TURBO Pascal library routines. You would
use a chain file when you want to write a Pascal program which
will be called to be run from within another Pascal program.
Can you see how the concept of "chaining" came about? One file
pulls another, just as one person could pull another by using a
real chain.
In order for a .CHN file to work, there must have been first a
regular .COM (OR .CMD) file loaded in memory.
The idea that you will have either a .COM file or a .CMD file
(depending upon your operating system) should be obvious by
now, so I am not going to keep mentioning both. From now on, I
will simply use the term .COM file and assume that you know
what I mean.
Find run-time error
Do you remember about how TURBO Pascal will find an error in
a program in memory when you try to compile it and then run it?
Do you remember how it will send you right back into the editor,
and show you where it thinks your mistake occurred?
6-12
Run Time Error
Well, one of the library routines included in any TURBO Pascal-

generated .COM file is similar to this. When you run a .COM file
and a mistake occurs, the program's execution will stop, and you
will be presented with a message like this:
Run-time errorOl, PC=lB56
Program aborted
This is telling you that a Run-time error occurred, and that the
computer has stopped executing your program. A run-time error
is one which happens while a program is being executed-or
running. The number (01, or whatever) refers to what TURBO
Pascal thinks the mistake was. You can look up the error, by its
number, in the back of your Reference Manual (if it is of interestto
you).
6-13
PC has to do with what is called the Program Counter; a thing in

the computer which keeps track of just where the computer is in
the computer program at any time. Its function is similar to your
following the street names on a map with your thumb, as you go
down an avenue. It is written in hexadecimal (base 16) math. (You
could translate it into decimal (base 10) math by using SideKick,
another one of my programs available from Borland International.)
When a PC number appears, you should make a quick note of it.

Later, when you have TURBO Pascal going, and the program in
question loaded, you can find the line where the error occurred.
You do so by getting into the compiler Options menu, pressing F
(for Find run-time error), and then answering with the PC number
you made a note of when you are prompted for it. The compiler
will point out where in the source program the error occurred.
Quit (the compiler Options menu)
Anytime you start something, you will eventually want to quit. It's
the same with this menu. The only difference is: when you press
Q to leave this menu, you will only go back to TURBO Pascal's
Main Menu.
Save
One of the sr:narter things you will do in programming, is to SAVE

ALL YOU DO!!!! Is that clear?
SAVE early and often. The reason for this is that sometime,
somewhere, when you least expect it, the power to your
computer will be interrupted. When that happens, all the work
you have done since your last save will go into never-never land.
Like the Good Book says, "A word to the wise, ... "
Oir
Every now and then, you will want to be reminded of the name of
the file you were working on the last time. If it is on the current
logged drive, pressing 0 will allow you to find out. Or you may
simply want to have a list of the files you have stored on disk.
Pressing 0 will show you those files as well.
6-14
What happens when you press D is that you are prompted for a
mask. If you wanted to see only the files ending in .PAS, for
instance, type in an asterisk (*) and a "dot-pas" like this:
*.pas <Enter>
All the files on the currently logged drive ending in .PAS will be listed for
you. If you want all the files to be listed, simply press:
<Return>
Remember, only the files on the logged drive (and the active directory)
will be listed. You will have to change the logged drive (see preSSing L,
above) before pressing 0 to get a listing of the directory, if it is different
from the one you are currently working with.
A hint: I usually press 0 to get a listing of the .PAS files on my drive.

Then, while they are still on the screen before me, I press W for work file
and enter the name of the file I want while I can see its name before me.
(If you know what wildcards are, you can use * to match any group of
characters or? to match any single character in the mask.)
eXecute
On some implementations of TURBO Pascal-that is "on some

kinds of computers running TURBO Pascal" -you are able to
run another (non-TURBO Pascal) program from within TURBO
Pascal. You do so by pressing X and then entering the name of
the program you want to run. When it is over (finished running)
you are automatically brought back into TURBO Pascal.
Quit
All good things have to come to an end, every now and then. To
leave TURBO Pascal to get back into your computer's disk
operating system (DOS), pressQ. If you haven't saved yourwork
file, you'll be asked:
Workflle MYNAME.PAS not saved Save (YIN)?
6-15
to which you should almost always reply Y.

Memory Space
The message at the bottom of the menu:

Text: 0 bytes
Free: nnnnn bytes
These are two more pieces of information available to you, and

which are best discussed together.
You are allowed from about 32 thousand characters (letters,
numbers, and symbols) less than your computer's memory
capacity in your program files. That means that if you have a 64K
computer, your program can be up to 32 thousand characters
long. If you have more memory, your programs can be substan-
tially larger. If that makes you nervous, this chapter has about
20,000 characters in it, so far. If your program gets so big that it
won't fit in the allowed space, you can break it into smaller, more
manageable, pieces which can then be cHained together. This
part ofthe menu, then, lets you know how long your source file is,
and how much room you have left to work with.
6.3 CONCLUSION
Well, that ends TURBO Pascal for the Absolute Novice. You're
now ready to start learning how to program in Pascal, which, of
course, is the point of the next part of the book, A Programmer's
Guide to TURBO Pascal. I'm happy to have been your guide; I
hope that you've enjoyed it as much asl have. You've done fine
up until now, and I'm sure you'll zip right through the rest of the
book. Me? Well, the dogs are chasing the burro (or is it the other
way around?), the chickens are out of their yard, and I've got a
few programming projects of my own that need to be finished, so
I'll see you around ...
6-16
PART II
A PROGRAMMER'S GUIDE
TO TURBO PASCAL
Part 2 Chapter 7 The Basics of Pascal
7. THE BASICS OF PASCAL
"Things are a/ways at their best in their beginning. "
-Blaise Pascal, Lettres Provinciales, no. 4

Okay. Either you skipped Part I, or you're a fast reader.
Whichever it is, you're aboutto get a quick introduction to Pascal.
First, you're going to learn how to write a simple Pascal program
by converting a simple BASIC program. After that, you'll learn
some Pascal terms which you'll be using throughout the rest of
the book. You'll undoubtedly have lots of questions as you go
along; rest assured that most (if not all) will have been answered
in great detail by the time you finish Part II.
7.1 A QUICK EXAMPLE
If you've gotten this far (and you obviously have, or else you
wouldn't be reading this), then you should know the fundamentals
of how to use TURBO Pascal. (That is, you should know how to
use the parts of the Main Menu; I don't yet expect that you can
write a program.) If my assumption is not correct, go back
through Part I again. Of course, you won't really need that for the
following example, since it's just a penciland-paper exercise
anyway, but you will need it before you go too much farther into
Part II.
Having gotten that out of the way, let's take a simple BASIC
program and convert it into a simple Pascal program. Suppose
our BASIC program looks like this:
100 REM SIMPLE BASIC PROGRAM
110 INPUTA
120 INPUTB
130 C=A+B
140 PRINTC
160 STOP
160 REM END OF SIMPLE BASIC PROGRAM
7-1
Chapter 7 The Basics of Pascal Part 2
This program allows you to enter two numbers, which are stored
in variables A and B. These two values are then added together
and stored in C. The contents of C are then printed out so that you
can see the sum of the two numbers that you entered. So far, so
good. Now comes the fun part: converting this innocuous little
piece of code into a real, live Pascal program. You're going to do
it a step at a time, so that you can understand what you're doing
and why. As usual, the explanation is more complicated than the
actual process.
The first step is to remove all the line numbers, since you don't
need them. Even if you wanted them, you would still remove
them, since Pascal uses other means of getting around than
jumping to a given line. Your program now looks like this:
REM SIMPLE BASIC PROGRAM
INPUT A
INPUTB
C=A+B
PRINTC
STOP
REM END OF SIMPLE BASIC PROGRAM
Next, you must change the names of the inputloutput (liD)

commands. Instead of INPUT, you will use the Pascal procedure
ReadLn, and you will replace PRINTwith the procedure WriteLn.
Now you have:
ReadLn(A)
ReadLn(B)
C=A+B
WriteLn(C)
STOP
Next, three more changes. First, you will replace the equals sign
(=) in the fourth line with a colon followed by an equal sign (:=).
Pascal uses := for assigning values and reserves the plain = for
comparing values.
7-2
Your second change is to put a semicolon (;) between all

executable statements. Since a Pascal statement can (and often
does) occupy several lines, the semicolon marks where one
statement stops and the next one begi ns. For neatness' sake,
you'll place the semicolon at the end of the precedi ng statement.
Lastly, you will eliminate the command STOP, which is not
needed or used in Pascal (and isn't really needed in your BASIC
version here).
Your program has changed a little more:
ReadLn(A);
ReadLn(B);
C:=A+B;
WriteLn(C)
Now for a major change. In BASIC, you can start execution on

any line, selectively execute groups of lines currently in memory,
and so on. In contrast, a PASCAL program is a well-defined
group of instructions, with what is known as the main body.
Execution always begins at the start of the main body and
proceeds through to the end. So you need to tell Pascal that your
program is a program and then place your instructions in the
main body. After you do this, it will look like this:
program Simple;
begin
Readln( A);
ReadLnCB);
C:=A+B;
WriteLn(C)
REM end OF SIMPLE BASIC PROGRAM
end
On to the next item. First some background: Pascal was designed
as a teaching language and, like most teachers, will pick nits (but
only for your own good, mind you). Case in point: BASIC allows
you to create variables as you go along, while Pascal demands
7-3
that all (and I do mean ALL) variables be declared before they can
be used. So let's declare our variables:
program simple;
var
A,B,C : Integer;
begin
REM SIMPLE BABIC PROGRAM
ReadLn(A);
ReadLn(B);
C:=A+B;
WriteLn(C)
REM end OF SIMPLE BABIC PROGRAM
end.
Onlyonething remains to makeyoursimple BASIC program into

a simple Pascal program: change the comment statements. In
BASIC, each comment line starts with REM (for REMINDER). In
Pascal, a comment starts with the character { and ends with the
character }, and you can have as many lines of comments
between those two characters as you want. If your terminal
doesn't have {and}, or if you just don't like curly braces, then you
can use the sequences (* and *) instead. You will now add a few
comments to ensure that your program is well documented. And
here's your final program:
program Simple;
{
A simple Pascal program converted from BABIC.
DATE: 17 June 1985
AUTHOR: Put your name here
var
A,B : Integer; { input variables }
C : Integer; { output variable }
begin { main body of program Simple
ReadLn(A);
ReadLn(B);
C:=A+B;
WriteLn(C)
end. { of program Simple
7-4
Here it is, an honest-to-goodness Pascal program, guaranteed to

compile and run under TURBO Pascal! It really doesn't look all
that different from your original BASIC program. And you
thought Pascal was hard!
Here's a review of the steps you used to translate your BASIC
program into Pascal:
1. Delete all the line numbers
2. Replace I/O statements
3. Replace assignment = with :=
4. Add semicolons at the end of statements
5. Drop STOP at the end of the program
6. Add program Name / begin ... end.
7. Declare all variables in the var section
8. Convert REM statements
It would be convenient if this set of rules was enough for all such
translations. Unfortunately, this isfarfrom true. In fact, these rules
will only translate a small subset of all BASIC programs.
Moreover, most BASIC programs over a certain size cannot be
translated at all, only rewritten. However, your purpose here is
not to show you how to convert BASIC programs into Pascal; it is
to show you that you can learn how to write simple Pascal
programs in just a few minutes.
7.2 SOME PASCAL TERMS

You're going to learn about some terms that will be used
throughout the book. Some of them (such as identifier) can be
fully explained before you leave this chapter. Others (such as
statement) will have their definitions expanded as you go through
the book, so don't get upset if ourfirst discussion doesn't seem to
be all-encompassing.
7-5
7.2.1 Characters
Let's start at bedrock. Your program will be composed of letters,

digits, spaces, and other printing characters. These are known as
ASCII characters (ASCII stands for American Standard Code for
Information Interchange, if you care. It is a computer-industry
standard way of representing all of the possible characters on
your keyboard).
There are 128 standard ASCII characters (numbered 0 though
127), and the table at the end of this chapter shows you all of
them. Notice that the first 32 have strange 2-and 3-letter abbrevia-
tions. These are known as control characters, because they're
used to control certain computer operations. For example, if your
program tried to write ASCII character 7 (BEL) on the screen, you
would hear a beep instead (on an old teletypewriter, you would
hear an actual bell ring, hence the name BEL, get it?).
Other commonly used control characters are BS (backspace),
CR (carriage return), and ESC (escape). You can type most of
these characters from your keyboard by typing the corresponding
character in the third column of the table (starting with character
64, "@" and going through character" _ ") while holding down
the Ctrl (or Control) key. What the computer (or, more accurately,
the program currently running in the computer) will do with them
is another question entirely.
Control characters are usualty written with Ctrl- or with an up-
arrow front. For example, character 7 (BEL) would be written:
Ctrl-Gor
By the way, there's another non-printing character in the ASCII

set, that is, one with a value greater than 31. See if you can find it
(Hint: it's not number 32-that's really a space; that is, the
character you get if you hit the space bar, which is actually a
printing character to a computer.)
7-6
Now on to the printing characters. As you can see from the table,
the ASCII character set contains all the letters of the alphabet, in
both upper and lower case. The upper case letters start at code 65
and go to 90, while the lower case ones start at 97 and go to 122. A
little math shows that the ASCII value of any lower case letter is
exactly 32 greater than 'its upper case equivalent. Remember
that-it can come in handy. The ASCII set also has all the decimal
digits (0 through 9), starting at code 48. Letters and digits
together are collectively known as alphanumeric characters.
Of course, there are plenty of visible characters besides just the
alphanumerics. Starting at 32 (which, as I said, is a space:" "), you
have several punctuation and other special characters in addition
to the letters and digits. These are usually called special
characters. The alphanumerics and the special characters togeth-
er are known as the printing characters, and they plus the control
characters form the entire ASCII character set.
Ta summarize: characters 0 through 31 are the control characters;
along with character 127 (did you find it?), they form the non-
printing characters. Characters 65 through 90 ('A' ... 'Z') and 97
through 122 ('a' ... 'z') are the letters, while characters 48 through
57 ('0' ... '9') are digits. Letters and digits together are known as
alphanumeric characters. All other characters not already men-
tioned make up the special characters.The alphanumeric and the
special characters together make up the pri nti ng characters. And
the printing and non-printing characters form the entire ASCII
character set.
Well, almost the entire set. Many computers recognize characters
128 through 255 as special printing or non-printing characters.
For example, the IBM PC uses those codes to represent foreign
characters, graphics characters, and special symbols. However,
since those are not standard for all makes of computers, they
really can't be listed here. Just be aware that you can have
characters with values greater that 127.
7-7
7.2.2 Identifiers
Okay, now that you completely understand ASCII characters,

you can learn about identifiers. In Pascal, many things have
names: programs, constants, data types, variables, procedures,
functions, you name it (if you'll pardon the expression). Accordi ng
to Standard Pascal, an identifier must start with a letter and is
then followed by zero or more letters and/or digits. TURBO
Pascal also allows you to use the underline character "_" (ASCII
95) anywhere you can use a letter, which means anywhere.
The case (capitalization) of letters is irrelevant, so that the
identifiers HOTSTUFF, HotStuff, and hotstuff are all the same.
Note, though that the underline character does make a difference;
hotstuff and hoLstuff are two different identifiers. As for length,
well, you can also make identifiers as long as you like (just about).
The actual limit is 127 characters (letters, numbers, and under-
scores), which should be long enough to please anyone.
7.2.3 Reserved Words
Pascal uses some special identifiers for putting programs togeth-

er. You used fourofthem in yourquick example above: program,
var, begin, and end.These identifiers (and the others listed below)
are called reserved words. That means you can't use them to
name programs, constants, data types, and so forth. For example,
you couldn't declare a variable named Program or call a program
Begin. You can only use them in the way Pascal (and Niklaus
Wirth) decrees. Here's a list of all the reserved words (sometimes
called keywords) in Standard Pascal:
aD4 8rrQ" begin case ccmst div
do downto else end. file for
fanction goto if in label mod
nil not of or packed procedure
program. record repeat set then to
type untU var while with
7-8
And here are some reserved words that TURBO Pascal uses as
well:
absolute external iDline 8hl 8hr

xor
If they all look strange now, don't worry; you'll know them well by
the time you finish this book.
7.2.4 Symbols
Many special characters are used by Pascal for various purposes.

You've already seen how the semicolon (;) separates statements,
the sequence colon-equals (:=) assigns values to variables, curly
braces ({ and}) delimit comments, and the period (.) signals the
end of the program. Other commonly used symbols are
* / + - arithmetic and set operations

= > < >= <= <> comparison operations
[] array and set delimiters
() function and procedure parameter lists
pointer and file operations
record field delimiter
Certain reserved words also function as symbols:
not and or xor shl shr logical operations

div mod integer arithmetic operations
in set inclusion
7.2.5 Constants
You often need to use a specific, fixed value of some sort when
writing a program. For example, if you were solving some
7-9
geometrical problems, you might want to use the value of Pi

(3.1415926 ... ). Such a value is called a constant, since it can only
have one value. (The opposite of a constant is a variable, which is
an identifier that can have more than one value. More about those
later.) A constant can be one of (at least) seven types:
TYPE EXAMPLES
Integer 3 0 -17382
Byte 3 0 255
Real 2.71828 0.00 -6.67E-22
Char 'A' '$' '0'
Boolean true false
set [0 .. 9] [] ['a' ,'r:, 'u']
string 'STARS.TXT' 'Amt in $'
Also, the reserved word NIL is the constant value of an unassigned

pointer. This probably means nothing to you now, and you can
safely forget it until Chapter 17, which talks all about such things.
Incidentally, the first five types shown are the standard predefined
data types of TURBO Pascal (more on those in Chapter 9). Sets
are a type of data structure; we'll cover these in Chapter 16.
Strings will be discussed in Chapter 13.
7.2.6 Variables
Sometimes you need to work with values that are unknown or

that might change while the program is executing. To do this, you
make up a name (identifier) and declare it to be of some data type.
For example, in the sample program above, you declared the
identifiers A, a, and C to be of type Integer. These identifiers are
known as variables (since their values can vary-get it?). Here
are some more examples of variables:
7-10
program SampleVariables; {
purpose shaw different types
of variables
var
Valuel,Value2,8um Integer;
Radius ,Circumference Real;
Selected,Done Boolean;
Answer,Initial Char;
begin
Valuel := 53; Value2 := 228;
Sum:= Valuel + Value2;
Radius:= 40.25;
Circumference:=2.0 * Radius * Pi; {Pi is predefined as 3.l4l5926536}
Selected:= True;Done := not Selected;
Answer: = 'Y' ;Initial : = 'd.'
end. { of program SampleVariables }
You can, of course, type in, compile, and run this program. Be
warned though: since there's no input or output (I/O, for short),
you won't see anything happen. A safe program, but dull.
7.2.7 Expressions
Just like constants and variables, expressions of different types

can be created. An expression of a given type contains constants
and/or variables of that same or compatible types, along with
appropriate symbols (if needed). When the program encounters
an expression, it evaluates it, that is, it combines all the constants
and/or variables and comes up with a single value of the
appropriate type.
In yourfirstsample program, you havethe integer expression A +
B. When the program evaluates this expression, it gets the value
of A and adds to it the value of B. The result issome integer value,
which is then stored in C. Likewise, in your second sample
program, you had the integer expression Value1 x Value2, and
the real expression 2.0 * Radius * Pi.
7-11
See, it's not as bad as you thought. Really, you've probably seen
many expressions before; in your science and math classes, they
were called formulas (or, if you want to be classical, formulae).
Things like the circumference of a circle orthe velocity of a falling
object with respect to time are just expressions (of the type Real,
usually). In fact, FORTRAN (one of the earliest programming
languages) stands for FORmula TRANslator. So, if it helps you,
just think formula every time you use the term expression.
7.3 REVIEW
Each of these topics deserves (and will get) more discussion. I

wanted to define enough basic terms to let you start talking about
Pascal itself... priming the pump, so to speak. You should now
have some grasp of what ASCII characters, identifiers, symbols,
constants, variables, and expressions are. If you do, you're ready
to go on to Chapter 8 and learn about program structure and
statements. If you don't, reread this chapter and go on to Chapter
8 anyway. It'll come with time.
One more item before we leave Chapter 7 is the ASCII table I
promised you several pages back.
7-12
ASCII Character Set Table
0 NUL 32 (space) 64 @ 96 \
1 SOH. 33 ! 65 A 97 a
2 STX 34 " 66 B 98 b
3 ETX 35 # 67 C 99 c
4 EDT 36 $ 68 D 100 d
5 ENQ 37 0/0 69 E 101 e
6 ACK 38 &
, 70 F 102 f
7 BEL 39 71 G 103 9
8 BS 40 ( 72 H 104 h
9 HT 41 ) 73 I 105 i
10 LF 42 * 74 J 106 j
11 VT 43 + 75 K 107 k
12 FF 44 , 76 L 108 I
.-
13 CR 45 77 M 109 m
14 SO 46 78 N 110 n
15 SI 47 / 79 0 111 0
16 DLE 48 0 80 P 112 P
17 DCl 49 1 81 Q 113 q
18 002 50 2 82 R ·114 r
19 DC3 51 3 83 S 115 s
20 DCA 52 4 84 T 116 t
21 NAK 53 5 85 U 117 u
22 SYN 54 6 86 V 118 v
23 ETB 55 7 87 W 119 w
24 CAN 56 8 88 X 120 x
25 EM 57 9 89 Y 121 Y
26 SUB 58 90 Z 122 z
27 ESC 59 ; 91 [ 123 {
28 FS 60 < 92 \ 124 I
29 GS 61 = 93 ] 125 }
30 BB 62 > 94 A 126 -
31 US 63 ? 95 - 127 DEL
7-13
Part 2 Chapter 8 Program Structure
8. PROGRAM STRUCTURE
Pascal programs are composed of statements. That sounds

simple enough, and it is, but it's a little like saying the English
language is composed of the 26 letters of the alphabet. This
statement is also true, but it's also a trifle simplistic. Statements
describe the actions to be performed by the computer. They can
be as simple-or as complex-as your knowledge, your needs,
and your creativity. You have to start somewhere, and it's best to
start simple and with the two sections essential to all Pascal
programs. They are:',
1 The declaration section
2 The program body
To show the basic structure of Pascal, you need to set up a simple
program:
program Simple;
var
, A,B,O : Integer;
begin
Write( 'Enter two numbers: ');
ReadLn(A); ReadLn(B);
C:=A+B;
WriteLn( 'The sum is ',0)
end. { of program Simple }
This program writes the prompt:

Enter two numbers:
on the screen. The program waits for two numbers to be entered,
then computes and prints to the screen
The sum is <sum>.
This sequence of events will occur only once for each execution
of the program.
8-1
Chapter 8 Program Structure Part 2
8.1 THE DECLARATION SECTION

The declaration section consists of a series of statements.
8.1.1 The PROGRAM Statement

The first line of your program is the program statement with the
identifier Simple. TURBO Pascal programs can (but do not need
to) begin with a program statement. This statement contains the
reserved word program followed by the name of your program.
(A reserved word is one that only Pascal can use; that is, you
couldn't name some variable program).
Standard Pascal requires a list of file parameters (input, output),
but TURBO Pascal just ignores such a list (as, indeed, it ignores
the whole statement), so we'll also ignore it from now on. A
semicolon separates the program statement from the rest of the
program. Here are more examples:
program BudgetAnalys1s;
programK;
program WithAVeryLongN"amelndeed;
8. 1.2 Declaration Statements

In your sample program, you used three variables: A, B, and C.
Variables are named locations in memory containing a value that
can be changed during program execution. You'll learn more
about them in Chapter 9, but for now just think of them as
shoeboxes or pigeonholes that hold numbers or other infor-
mation.
Pascal demands that all (and that means al/) variables must be
declared before they can be used. Declared means that each
variable must have its name (A,B,C, for example) and type
(Integer, for example) defined before it is used. In fact, this is such
8-2
a key element of Pascal that you need to call special attention to

it. So memorize the following rule:
THE GREAT UNDERLYING RULE OF PASCAL:
All identifiers must be declared before they are used.
This rule (which I'll call The GURP, for short), applies to all
identifiers: constants, data types, variables, subprograms, and
any other use of an identifier. Right now, though, you're only
concerned with variables.
Variable declaration is done between the program statement
(program Simple;) and the main body statement (begin ... end.).
To signal the start of the variable declaration statements, you use
the reserved word var, then name your variables and give their
type(s). Each statement in this section contains a list of one or
more variable names, followed by a colon, then the variable type,
and ending with a semicolon. Given these requirements, Pascal
allows you flexibility in formatting these sections. It would be
equally correct to declare your variables this way:
var
A,B,C : Integer;
or
var
A Integer;
B Integer;
C Integer;
As mentioned back in Chapter 7, there are five predefined data
types in Pascal:lnteger, Byte, Real, Char, and Boolean. The first
four are just what they sound like:integer numbers (0,5, -18232);
byte values (0 through 255); real or floating point numbers (0.0,
3.14159, -35.232e-5); and ASCII characters ('A','-','3'). Boolean
variables have either of two values- True or False-and are used
to store the truth value of some proposition so that it can be tested
later .Here are some variable declarations for these different data
types:
8-3
var
Alive,Breathing,ConscioUB Boolean;
Age,Height,Weight Integer;
Score,Tries Byte;
Ratio),ercentage Real;
First,Middle,Last Char;
Note that a declaration statement (or any other kind of statement)

doesn't have to fit on just one line. If you wanted to be expansive,
you could rewrite this as:
var
Alive,
Breathing,
Conscious
: Boolean;
Age,
Height,
Weight
: Integer;
Score,
Tries
: Byte;
Ratio,
Percentage
: Real;
First,
Middle,
Last
: Char;
You would have lots of room to put a comment after each variable
name. You would also have very long programs. Here's yet
another variation:
var Alive : Boolean; Breathing : Boolean; Conscious : Boolean;
Age : Integer; Height : Integer; Weight : Integer;
Ratio : Real; Percentage : Real; First : Char;
Middle : Char; Last : Char;
8.2 THE PROGRAM BODY

The main body of a Pascal program consists of the statement
begin .•. end, with any number of executable statements between
these words. Execution always starts with the first statement after
8-4
begin and proceeds to the last statement before end (or, at least,
tries to). Notice that there may be several such begin...end pairs,
and that only the final end is followed by a period (.). This is the
only place where a period should follow end.
Between the reserved words begin and end in our sample
program are five executable statements:
Wr1te('Entertwo numbers: ');
ReadLn(A)j
ReadLn(B)j
C:=A+Bj
Wr1teLn( 'The sum is ',C)
Note that a semicolon appears at the end of the first four

statements. Pascal doesn't use lines to distinguish between
statements-just semicolons. More precisely, Pascal uses semi-
colons to separate statements.You could have written your
program this way:
program Simple j var A,B,C : Integer
; begbi Wr1te('Enter two numbers: ') j ReadLn(A)
j ReadLn(B) j C:=A+B ; Wr1teLn('The sum is ',C) end.
or this way:
program SIMPLE
var
A,B,C : Integer
,
begin
Wr1te('Enter two numbers: ')
,
ReadLn(A)
,
ReadLn(B)
,
C:=A+B
,
Wr1teLn('The sum is ',C)
end.
8-5
As you can see, semicolons are only used when it's necessary to
show where one statement ends and the next begins. You can
also see that Pascal also doesn't care about line indentation,
upper and lower case, and (sometimes) spaces. You have the
freedom to format your programs in a variety of ways. The best
policy is to pick a consistent, readable, and easy-to-use format,
then stick with it. The two alternate styles shown are neither
readable nor desirable to use, so we'll stick with the regular
TURBO format.
Now let's look at the five executable statements, one at a time.
Write('Enter two numbers: ')j
The command Write is really a predefined procedure telling your

program to write whatever is enclosed in the parentheses out to
the screen. It's one of a set of standard procedures and functions
that Pascal defines for you. Two tables at the end of this chapter
list the predefined procedures and functions that TU RBO Pascal
provides. You'll meet most of these as you tackle the appropriate
topics in this book.
As mentioned, Write will print its contents on the screen. (It will
do much more that that, as you will see in Chapter 18.)ln this
example, you have text -a string constant. Upon execution of this
statement, the message:
Enter two numbers:
will appear on the screen. The program will then wait with the
cursor two spaces away from the colon. It's waiting because the
next statement is
ReadLn(A)j
As you can guess, ReadLn is another standard procedure that

causes the program to wait for input from the keyboard (and
more; again, see Chapter 18). Here, since A is an Integer variable,
it's waiting for you to enter an Integer number. When you type in a
number, then press the Enter key, the program assigns that
number to the variable A in the parentheses. After you do this, the
program will wait again, because the next statement is
8-6
ReadLn(B);
It's waiting for another number, followed by Enter. Since the
program is waiting for Enter, you can correct any mistakes you
might have made in typing the number when you do so before
pressing Enter. Your program won't accept (read) data until
Enter is pressed.
Your program executes the next two statements on its own. With
the statement
C:=A+B;
the program evaluates the expression A + B. In other words, it

adds the contents (value) of A with that of B. It then gives that
value to the variable C. This is known as an assignment
statement, since it assigns some value to a variable. Let's suppose
that, when prompted by the program, you had entered the values
21 (for A) and 35 (for B).
After execution of this statement, the variable C would contain

the value 56. Note that Pascal uses := to assign the value on the
right to the variable on the left. By contrast, Pascal uses = to
compare two values (I F A =B THEN ... ). Be careful not to confuse
these two types of statements. You'll learn more about compari-
sons in Chapter 10.
After the assignment statement, we come to our last standard

procedure call:
WriteLn('The sum is ',C);

This prints on the screen the message
The sum is
followed by the value of C. For example, if you had indeed

entered 21 and 35 for A and B, then the message would be
The sum is 56
8-7
Since you used WriteLn, the cursor will be moved down to the
start of the next line after the message has been written.
The procedures Write and WriteLn will print out any number or
combination of messages and values enclosed in the parentheses.
Expressions (formulas) can be included as well. You could
eliminate the assignment statement and get the same message
by changing the last statement to be
WriteLn('The sum is ',A + B);
All separate messages and values within the parentheses must be
separated by commas.
8.3 COMMENTS
At this stage, you need to add one more element to your

program: comment statements, to ensure that your program is
well documented. A comment statement starts with a left curly
bracket ({) or the seq uence {*. It then conti nues for as many Ii nes
as you wish. It ends a right curly bracket (}) or the sequence *),
depending upon how you started it. Everything within the
comment statement is ignored. (Well, almost everything. See the
discussion of compiler options in the TURBO Pascal Reference
Manual. ) Here's our simple, yet well-documented program:
program Simple;
{
purpose adds two user-entered values and
prints out the sum
last update 21 April 1985
var {variable declaration statements }

A,B : Integer; { input variables }
o : Integer; { output variable }
{ main body of program Simple }
begin
Write(,Enter two numbers: ');
ReadLn(A);
ReadLn(B);
O:=A+B;
WriteLn('The sum is ',0);
end { of program Simple }
8-8
Noticethat you didn't haveto put a comment symbol on each line

of the program description between the program statement and
the variable declarations. I nstead, you just put the start comment
symbol ({), went on for as long as you wanted, then finished with
the end comment symbol (}).
One thing you have to really watch for is nested comments. Let's
say that you've written the following section of code:
for Indx := 1 to SysMax do
with Starmap'" [Indx] do begin
Write('SYSTEM: ',Name); { write system name }
Write(' Populatio~ ',Pop); { show population }
if Indx = CurStat.Loc.System { check location }
then WriteC' <current location> ');
WriteLn
end;
Don't worry if you don't understand the code; just notice the
comments to the right of three statements. They're a little
redundant, but they'll help illustrate the problem. Now, suppose
that you wanted to temporarily remove the Write statement
showing the population as well as the if ... then statement; that is,
you didn't want to delete them from the program, but you didn't
want them to execute, either. You could simply put comment
symbols before and after the statements:
for Indx := 1 to SysMax do
temporarily cut from program
Write(' Populatio~ ',Pop); { show population }
then WriteC' <current location> ');
WriteLn
end;
Your problem is solved, right? Wrong. The comment will start
o.k., but it will end when it finds the} after the words show
8-9
population. The if. .. then statement will be as "visible" as it was

before, and the following}, trying to end a comment that's already
done, becomes an illegal symbol. And, if you simply deleted the
comment { show population }, the same problem would crop
up with the comment { check location }.
There are two ways of handling this problem. The messy way is to
do something about any comments inside the one you're
creating (known as nested comments). You could delete them,
but that means you would have to re-enter them when you
"uncommented" those statements. That would be easy in this
case, but what if you needed to "comment out" a large section of
the program? Deleting and then re-entering all those comments
(assuming, of course, there are any) could be tedious. Another,
similar solution would be to change the end comment symbol (})
in each of those comments to something else. For example, you
could replace them with - (tilde):
{ tempora.:ri.1y cut from program
Write(' Fopuln"tion.; ,,Fop); { show population -
if Indx = CurStat.Loc.System { check location -
then Write(' <current location> ');
This works fine, but you must be sure to change each and every
one back when the code is "un-commented." Why? Because if
you miss one, the comment will merrily continue throughout
your pt'c,rFn.m until it finds an }.
An en r'. . :·:!':is can cause real headaches. If you're lucky, it will
mess ur.~ ____ .. ; program enough so that it won't compile, and you'll
be forced to find the error immediately. If you're not, the program
will compile just fine, and you won't see any problem until you
execute it. If you're really unlucky, the program will appear to
work just fine, and you won't discover that there's a problem until
after the first 5000 copies have been shipped.
There is a cleaner way of handling this problem, however.
Remember, you can use two different sets of delimiters for
comments: {,} and (*,*). What's more, TURBO Pascal allows you
to nest one kind of comment within the other. Allow isn't quite the
8-10
right word; it's a natural result of the comment definition. If you

start a comment with {, it will ignore everything it comes across,
including (* and *), until it finds}. The reverse is true with (*; that
is, it will ignore { and }.
The solution, then, is to consistently use one type of comment
-say, {,} - for all program comments and use the other type - in
this case, (*,*) - for blocking out sections of code. Your example
would now look like this:
for Indx := 1 to SysMa.x do

(*temporarily cut from program
Write(' Population: ',Pop); { show population
WriteLn
end;
A few more words about comments. Not only can they extend
across many lines, they can be stuck in anywhere. As a
demonstration of this, we present your example program with
comments stuck in everywhere. I'm not sure why you would want
to do this to a program, but it's good to have the freedom if you
ever need it.
program Simple{ file parms: }(input,output);
var{ iables}
A{ input} ,O{ input} ,O{ output} : Integer;
begin { main body of program }
Write{ to screen }(,Enter two numbers: ');
o {sum} := A {1st value} + B {2nd value};
WriteLn{to screen}('The sum 18 ',0)
end. { of pro( * bye! *) gram Simple }
Well, almost anywhere. Had you inserted comments in the strings
'Enter two numbers:' and 'The sum is " they wouldn't have been
comments. Instead, they would have been printed to the screen
along with the other text within the single quote marks.
8-11
You now know enough to start creating your own Pascal

programs. A quick reading of the next two chapters, especially
the first section of each, will give you even more information to
play around with. Feel free to do so; you can always come back to
this section later.
8.4 ADVANCED PROGRAM STRUCTURE
Program structure in Standard Pascal follows a definite format.

Some sections can be omitted, but no rearranging is allowed.
However, rearranging is allowed by TURBO Pascal. Here is the
complete format of a Standard Pascal program:
program Name( file parameters);
label
< label declarations>
canst
< const declarations>
type
< type declarations >
var
< variable declarations >
< subprograms >
begin
< main body of program >
end.
TURBO Pascal is more flexible than Standard Pascal in its

program structure. The label, const, type, and var sections can
be placed in any order and can occur more than once. They also
can be mixed with subprograms. This allows logical grouping of
definitions, such as:
8-12
const
XMax = 8;
YMax = 10;
ShipMax = 15;
type
XRange = l..XMax;
YRange = 1..YMax;
var
Sector : array [XRange,YRange ] of O..ShipMax;
const
A second advantage of this freedom is the ability to define typed

constants, which will be discussed below.
As you've seen, a program starts with the program statement.

The program statement is followed by the declaration section.
Here you define any labels, constants, data types, and/or
variables which will be used throughout the program. Each
section begins with the appropriate reserved word (Iabel,and so
forth), followed by one or more corresponding declarations.
Here's a brief description of each kind:
label one or more unsigned integers (0 .. 9999) or
identifiers, separated by commas, with the whole
list ended with a semicolon:
label
10,ProcExit,4213,Error;
const an identifier and a constant val ue, separated by

an equals sign (=) and ended with a semicolon;
also, a typed constant (see below):
const
XMax =8;
Yes = True;
Answer ='Y';
G = 6.673E-08;
8-13
type an identifier and a type definition, separated by

an equals sign (=) and ended with a semicolon:
type
XRange = 1.XMax;
language = (BASIC,FDRTRAN,Ada,Pascal,
FDRTH,C,COBOL,.ALGOL);
var one or more identifiers separated by commas,

followed by a colon and a type identifier or type
definition, and ended with (you guessed it!) a
semicolon:
var
A,B,C : Integer;
Flag : Boolean;
Choice : language;
X :XRange;
The const-type-var sequence allows definitions to cascade

down, so that you can change an entire program by merely
changing a few constants and recompiling. Combined with the
ability to repeat that sequence, this gives you great power and
flexibility in defining your data structures.
8.4. 1 Typed Constants

TURBO Pascal lets you declare what it calls typed constants.
These aren't really constants, though they are declared in the
canst section. Instead, they are pre-initialized variables. Suppose
you had two flags, Flag1 and Flag2, that you were going to use
throughout your program. Flag1 needed an initial value of True,
while Flag2 needed one of False. You might do the following:
8-14
program Whatever;
var
Flagl,Flag2 : Boolean;
begin
Flagl := True;
Flag2 := False;
end
This is fine, but if you have many such variables, it can consume
space and time to do all of the initialization. With TURBO Pascal,
you could instead do this:
program Whatever;
const
Flagl : Boolean = True;
Flag2 : Boolean = False;
begin
end
Flag1 and Flag2 can still be used as variables; that is, you can still
assign values to them through the course of your program.
However, they will now start out with the desired values and will
not have to be explicitly initialized.
Forthoseofyou who know something about Pascal, yes, you can
have typed constants for data structures such as arrays, strings,
sets, and records. All the details can be found in the TURBO
Pascal Reference Manual, Chapter 13.
8.4.2 Subprograms
Between the definitions and the main body of the program you
place any subprograms (procedures and functions) that you
might define. I'll talk more about these in Chapter 11, but a brief
explanation isn't out of place. A subprogram has exactly the
same structure as a program, except that it starts with a
8-15
procedure or function statement and ends with a semicolon

(instead of a period). A subprogram is executed just by men-
tioning its name along with any parameters it might take. Pascal
includes a set of predefined subprograms; that is, procedures
and functions that already exist without you having to define
them. You sa\y some in the example earlier in the chapter: Write,
WriteLn, and ReadLn. The two tables at the end of this chapter list
the standard subprograms of TURBO Pascal. As a quick example,
here's our sample program with the main body converted into a
procedure named CalculateSum:
program Sample;
{
purpose to demonstrate a subprogram
last update 28 August 1985
}
procedure CalculateSum;
{
purpose read in two values and print the sum
last update 28 August 1985
var
A,B,C : Integer;
begin { main body of procedure CalculateSum }
Write('Enter two numbers: ');
C :=A+ B;
WriteLn('The sum is ',C)
end; { of proc CalculateSum }
begin { main body of program Sample }

CalculateSum
end { of program Sample }
When the program runs, it only has one statement to execute:

CalculateSum. That tells it to execute the statements in the main
body of the procedure CalculateSum. When it's all done, it goes
back to the main body and, having no more statements, stops.
You can do an awful lot with subprograms, but I won't tell you
what until Chapter 11.
8-16
8.4.3 Block Statements
Following all the subprograms is the main body of the program

itself, consisting of the reserved words begin and end, with some
number of statements between them. This is just a special case of
what is known as a block statement. A block statement has the
sequence begin ... end, with zero or more statements (separated
by semicolons!) in between. A block statement can be used
anywhere a regular statement can be used, and several places
where a regular statement can't. For example, the main body of a
program is just a block statement followed by a period, while the
main body of a subprogram is a block statement followed by a
semicolon. The real"advantage of block statements shows up in
control structures (see Chapter 10). For example, the if••• then
statement is defined as
if <Boolean expression>
then <statement>
But you often wish to execute more than one statement when a
given condition is true. Solution:use a block statement, like this:
if Max < Min then begin { force min < = max }
temp := max;
max := min;
min := temp
end;
You also used a block statement back in the section on

comments. You were using awith statement, which allows you to
easily see the fields of a record data type. It has the format:
with <identifier(s» do <statement>
The statement here was the block statement:
begin
WriteC'SYSTEM: ',Narne);
Write(' Population; ',Pop);
if Indx = CurStat.Loc.System
WriteLn
end;
8-17
You can nest block statements; that is, have block statements
within block statements. For example, you could have a block
statement in the if ... then statement, yielding:
with StarmapA [Indx] do begin
Write('SYSTEM: ',Name);
Write(' Population; ',Pop);
if Indx = CurStat.Loc.System then begin
Write(' <current location> ');
LocationFound := True
end;
WriteLn
end;
Simple, huh? You'll use block statements a lot throughout the

book, especially after Chapter 10.
That's all for advanced program structures for now. Before
continuing to the next chapter, here are the two tables I promised
you back in the early parts of this chapter. First is a table of
predefined procedures, followed by a table of predefined
functions.
PREDEFINED PROCEDURES
Assign(F,N) Assign file F to file name N

BOos(R) make BOOS call with regs R (CP/M-86)
BOos(F,P) make BOOS call to F with P (CP/M)
Bios(F,P) make BIOS call to F with P (CP/M)
BlockRead(F,O,N) read N blocks from 0 to file F
BlockWrite(F,O,N) write N blocks from 0 to file F
Chain(F) chain to file F
Close(F) close file F
ClrEol clear to end of current screen line
ClrScr clear entire screen
CrtExit send terminal reset string
Crtlnit send terminal init string
Oelay(M) delay M milliseconds
Oelete(S,P,L) delete section of string S
OelLine delete current screen line
Oispose(P) recover memory used by PA
Erase(F) delete file F
8-18
Execute(F) execute file F

FiIiChar(V,L,D) fill V with data 0 for L bytes
GetMem(P,I) allocate I bytes of RAM for p-"
GotoXY (X, Y) move cursor to X,Y (1,1=upper left)
Halt stop program execution
HighVideo sets high intensity display mode
Insert(S,D,P) insert string 0 into S
InsLine insert line on screen
LongSeek(F,P) special Seek routine (MS-DOS)
LowVideo switch to "dim" video output
Mark(P) mark heap pointer at p-"
Move(S,D,L) move L bytes from S to 0
MSDos(R) make call to MS-DOS with regs R
(MS-DOS)
New(P) create memory for p-"
NormVideo switch to normal video output from dim
Randomize init random seed (see Chapter 20)
Read(P1, ... ) read items in from keyboard
Read(F,P1 ... ) read items from file F
ReadLn(P1, ... ) as "read", but moves to new line at end
ReadLn(F,P1 ... ) ditto, but from file F (textfile only)
Release(P) reset heap pointer to p-"
Rename(F) rename file F
Reset(F) open file for input
Rewrite(F) open file for output
Seek(F,P) move to record Pin fi Ie F
Str(N,S) convert number N to string S
Val(S,N,P) convert string S to number N (error at P)
Write(P1, ... ) write items out to screen
Write(F,P1 ... ) write items out to file F
WriteLn(P1, .. ) as "write", but starts new line at end
WriteLn(F,P1 ... ) ditto, but to file F (textfile only)
PREDEFINED FUNCTIONS
Abs(A) absolute value of A (Real, Integer)
Addr(V) address of variable V (Pointer)
Addr(SubP) address of subprogram SubP (Pointer)
ArcTan(X) arctangent of X (Real)
BDos(F,P) does BOOS call; returns reg A (CP/M)
BDosHL(F,P) ditto; returns HL pair (CP/M)
8-19
Bios(F,P), , does BIOS call; returns reg A (CP/M)

~iosHL(F,P) ditto; returns HL pair (CP/M)
Chr(l) qharact~r with ASCII. value I (Char)
Concat(S, .. ,S) concatenation of strings (String)
Copy(S,P,L) substring at P length L (String)
Cos(X) cosine of X (R,sal)
EOF(F) end-of-file,test on file F (Boolean)
EOLn(F) end-of-line test on file F (Boolean)
E?<p(X) exponential of X (Real)
FilePos(F) current record inJile F (Integer)
FHeSize(F) total records 'in file F (Integer)
Frac(X) fractional portion of X (Real)
Hi(l) upper byte of I (Integer)
Int(X) integer portion of X (Real)
KeyPressed keyboard status flag (Boolean)
Length(S) length of ~tringS (Integer)
Ln(X) natu~al logarithm of X (Real)
Lq(l) lower byte of I (I nteger)
MemAvail bytes/paragraphs available (Integer)
Pos(P,S) position of str P in str S (Integer)
Ptr(l) poiqter to address I (Pointer)
O'dd(l) odd/even test of I (Boolean)
Ord(Sc) ordinal value of scalar variable (Integer)
Pred(Sc) predecessor of scalar value (same type)
Random' random value from 0.0 to 0.999... (Real)
Random(l) random value from 0 to 1-1 (Integer)
Round(X) rounded;..off value of Real (Integer)
ShL(I) Shift left (Integer)
ShR(I) Shift right (l'nteger)
$in(X) sine of X (Real)
$izeOf(V) size in bytes of variable V (Integer)
SizeOf(T) size in bytes of data type T (Integer)
Sqr(A) A ~ A (Real, Integer)
Sqrt(A) square roo~ of A (Real)
Succ(Sc) successor of scalar value (same type)
Swap(l) I with, upper, lower bytes swapped
(Integer)
Trunc(X) truncated value of X (Integer)
UpCase(C) C converted to uppercase (char)
8-20
Part 2 Chapter 9 Predefined Data Types
9. PREDEFINED DATA TYPES
I n the fi rst two chapters of th is part of the book, we've referred to

constants, variables, expressions, and data types. Well, now I am
going to teach you more about them. Let's start with variables.
9.1 VARIABLES
Think of Tupperware®. Think of pages and pages of different

plastic containers, designed to hold a variety of items. Some are
designed for liquids, others for dry goods, yet others for lunch
meats. Similar or identical containers can hold similar or
identical items. Items can be transferred from one container to
another. '
In much the same way, you can use variables in your programs.
You can declare lots and lots of variables, designed to hold
different kinds of information. Some are designed for numbers,
others for characters, yet others for logical values. Similar or
identical types of variables can hold similar or identical values.
One variable can receive the value of another.'
Actually, the Tupperware® analogy is weak in a few spots, but it
should convey the basic idea:variables hold stuff for later use.
And, instead of buying them at a home party, you get create
variables by declaring them. Remember our sample program?
program Sample;
var
A,B,C, : Integer;
begin
C:=A+B;
Write('The sum is ',C)
en4.{ of program Sample}
9-1
Chapter 9 Predefined Data Types Part 2
For your program, you needed three variables: two to hold the
values you typed in, and one to hold the sum of the other two. So
you created them by declaring them. You declared them by (1)
listing their names (identifiers), and (2) saying what type of
variables they were (in this case, Integer). You did all this in the
variable declaration section, which started with the reserved
word var and ended when you reached the main body of the
program. Here you get back to the Great Underlying Rule of
Pascal (or, GURP): All identifiers must be declared before they
are used. In your case, it means that A, B, and C must all be
declared before you can use them in your program. Suppose you
added a variable to your program:
program Sample;
var
A,B,O : Integer;
begin
ReadLn(A);
ReadLn(B);
O:=A+B;
D:=A-B;
WriteLn('The sum is ',0);
WriteLn( 'The difference is ',D)
end.. { of program Sample}
Question: is this an acceptable Pascal program? Answer: no!

Why not? Because you're using an identifier, 0, which you
haven't defined. If you were to compile this program, you'd get an
error stating "Unknown identifier" when the compiler reached
the statement 0:= A -B. How do you fix it? Like this:
program Sample;
var
A,B,O,D : Integer;
begin
Write( 'Enter two numbers: ');
ReadLn(A);
ReadLn(B);
O:=A+B;
D := A -B; WriteLn('The sum is ',0);
WriteLn( 'The difference is ',D)
9-2
Now 0 has been properly defined as an integer variable, the same

as the others used. But why did you decide to make them all of
type I nteger? And just what does that mean, anyway? Let's look at
the four basic data types of Pascal so see if you can get an answer.
9.2 INTEGER
You know what integers are: they're counting numbers (0, 1, 2,

3, ... ) with negative numbers (-1, -2, -3, ... ) thrown in. The largest
integer you can work with depends upon the amount of space
used for integer variables. TURBO Pascal allocates two bytes,
giving a maximum positive value of 32767 (known as Maxlnt).
The maximum negative value is -(Maxlnt+1). So, TURBO Pascal
allows the following range of values for integers:
-38768 -38767 -38766 . . . 0 . . . 38768 38766 38767
Remember your discussion of constants and expressions? An

integer constant is simply a string of digits (no commas, please)
with an optional - (minus sign) in front. An integer expression
(formula) contains integer constants and/or integer variables
and/or integer operators. The result of an integer expression is
some integer value. Your sample program has two such
expressions:
a+b add the values of a and b
a- b subtract the value of b from that of a
You usually find expressions on the right side of assignment
statements (as in your sample program). The program evaluates
the expression using the current contents of any variables
involved and then stores the resulting value in the variable on the
left side of the :=. The integer operators that can appear in an
expression are:
= addition
subtraction (and unary minus)
* multiplication
div division
mod remainder
9-3
Addition, subtraction, and multiplication all work just like you'd

expect them to. The rest may need a little explaining, though.
Let's start with div and mod. Remember when you first learned
division? When you divided 4 into 11, you didn't get 2.75. You got
2 (the quotient) with a remainder of 3. Why? You were doing
integer division, which required that the' results also be integers.
Well, that's what div and mod do:A div B returns the quotient,
while A mod B returns the remainder. Maybe the following
examples will help:
8 div4 =2 8mod4=0
9 div4=2 9 mod 4 = 1
10 div4 = 2 10 mod 4 =2
11 div4 = 2 11 mod 4 =3
12 div4 =3 12 mod 4 =0
Two other things to keep in mind about div and mod. First, if A is
less than B, then A div B will equal O.ln other words, while 200 div
200 equals 1, 199 div 200 equals O. Second, if B equals 0, then
both A div B and A mod B will result in some sort of error when
they are executed. Make sure that you keep both situations in
mind when working with these integer operations.
O.K., you've got integer constants, integer variables, and integer
operators, and you can combine these into integer expressions.
You're all done, right? Well, consider the following program:
program Sample;
var
A,B,O : Integer;
begin
A:= 10;
B:=5;
0:= 5 * A + 2 * B;
Write('O = ',0)
Question: what value will this program print out? Well, that
depends upon how you evaluate the integer expression 5 *A x 2 *
B. Here are four different ways you could evaluate it (remember,
A=10andB=5):
9-4
((6 * 10) + 2) * 6 =260

(6 * 10) + (2 * 5) =60
5 * (10 + 2) * 5 =300
5 * (10 + (2 * 5)) = 100
As you can see, the order of evaluation can make a big difference.
Which brings you back to your question:What will Pascal do with
this expression? Answer: it will use operator precedence. Pascal
will carry out all multiply, divide, and remainder operations before
any addition or subtraction. A complete list of operator prec-
edence is given in the first of two tables at the end of this chapter.
So the correct (according to Pascal) evaluation is:
(5 * 10) + (2 * 5) = 50 + 10 = 60
Having solved that problem, let's tackle the next one. How would
you evaluate the expression A div 2 * B? You have two choices
(again, A = 10 and B = 5):
(10 div 2) * 5 = 25
10 div (2 * 5) =1
You would probably choose the first one, and you'd be right.
When operators of equal precedence show up, you take them in
the order they come: left to right.
Next question: what if you don't like the order of evaluation?

What if you wanted 5 * A + 2 * B to equal 260 instead of 60?
Answer:do just what you did above, namely, use parentheses.
Any part of an expression within parentheses is evaluated before
it's combined with anything outside of the parentheses. This
applies to nested parentheses, as well. Our examples above, well
worth repeating, show how this works:
((6*10)+2)*5 (50+2)*5 =52*5 260
5*(10+2)*5 5*12*5 =60*5 300
(5*10)+(2*5) 50+10 60
5*(10+(2*5)) 5*(10+10) =5*20 100
9-5
So, assuming that you do want an answer of 260, you have to

modify your program like this:
program Sample;
var
A,B,C : Integer;
begin
A:= 10;
B :=5;
C := ((5 * A) + 2) * B;
Write('C = ',C)
end. { of program Sample}
Actually, since multiplication has precedence over addition, you

could drop the innermost set of parentheses and write:
C := (5 *A + 2) * B;
What started out as a short discussion of integer expressions has
gotten rather long. And you're not done yet. Let's modify your
program a little more:
program Sample;
var
A,B,C : Integer;
begin
A:= 100;
B :=200;
C : = ((5 * A) + 2) * B;
Write('C = ' ,C)
end. { of program Sample}
What value will you get when this program runs? Well, let's look at
it:
((5*100)+2)*200 = (500+2)*200 = 502*200 = 100400

The variable C, being an integer, can only hold values up to
Maxlnt, which equals 32767. You've caused a value range error.
This simply means that you've tried to give a variable some value
outside of its allowable range. What wi II happen? The variable C
will end up with the lower 16 bits of the result, which in this case
will come out as -30672. So, you need to BE CAREFUL. The
9-6
part 2 Chapter 9 Predefined Data Types
worst kind of program bugs are those that allow your program to
keep running with values that no longer have any relation to your
original data. A few simple steps can check the result to see if it is
within the realm of realism.
I n addition to the operators described above, Pascal offers a few
predefined functions that work on integers. They're listed below,
with the data type of the result given in parentheses:
Abs(l) returns absolute value of I (integer)
Odd(l) returns true if I is odd (boolean)
Pred(l) returns 1-1 (integer)
Random(l) returns random integer from 0 to 1-1
Sqr(l) returns 1*1 (integer)
Sqrt(l) returns square root of I (real)
Succ(l) returns 1+1 (integer)
9.2.1 Integers as Unsigned Values
Sometimes you need to use integers as unsigned 16-bit values.

The most common reason is to represent some address in
memory (RAM). There is, however, a problem you run into: once
you get above 32K, you have to use a negative number. The
sequence goes like this:
0, I, 2, ... , 32766, 32767, -32768, -32767, ... , -2, -1
Hardly convenient or intuitive. TURBO Pascal, however, gives

you a way around this:hexadecimal (base 16) constants. A hex
value can have up to four digits, where each digit can be 0
through 9 or A (=10) through F (=15). All you haveto do is start the
constant with a dollar sign ( $ ). Here, for example, is the same
sequence expressed as hex constants:
$0, $1, $2, ... , $7FFE, $7FFF, $8000, $8001, ... , $FFFE, $FFFF
Since memory addresses are often expressed in hexadecimal

anyway, this makes for easy, convenient manipulation of address
values.
9-7
TURBO Pascal provides other ways of treating integers as

unsigned values. In addition to the operators and functions
described above, TURBO provide six operators, two functions,
and a procedure that let you do bit and byte manipulation on
integer values. First, the operators:
not logical negation
and logical AND
or logical OR
xor logical exclusive OR
shl shift left
shr shift right
The first operator, not, is a unary operator; that means that it only
affects the integer value it precedes. It differs from unary minus
(-) in that it flips each bit within the value from 0 to 1 or 1 to 0, while
a - does an arithmetic negation. An example using hex constants
may make things clear:
- $0001 = $FFFF (-1)
~ $0001 = $FFFE (-2)
The unary minus changed 1 to -1, which has a hex value of

$FFFF. The NOT operator flipped each bit in $0001, which means
that it changed all bits from O's to 1's except for the very last,
which changed from a 1 to a O. The result-$FFFE-happens to
be equivalent to a signed value of -2.
The next three operators perform the standard logical functions
defined in the second of two tables at the end of this chapter.
They can be used to combine or mask values. For example, if you
want to see if a particular bit is set (=1), you could do something
like this:
var
BitBet Boolean;
Val,Mask Integer;
begin
BitBet := (Val and Mask) <> 0;
9-8
The Boolean variable BitSet will be set to True if any of the bits in
Val match any of the bits in Mask. By the same token, the
statement:
Val:= Val or Mask;
will cause each bit in Val to be set to 1 if the corresponding bit in

Mask is also 1.
The last two operators, shl and shr, allow you to shift all the bits
within an integer to the left or right, respectively. The first
operand is the value to be shifted; the second, how many places
to shift. Suppose that we set Val := $0810, which has a bit pattern
of (0000,1000,0001,0000). Here are some expressions with their
resulting values:
Valshl1 $1020 (0001,0000,0010,0000 )
Valshr 1 $0408 (0000,0100,0000,1000 )
Val shl 6 $0200 ( 0000,0010,0000,0000)
Val shr 4 $0081 (0000,0000,1000,0001 )
As you can see, shl and shr shift in a's and throwaway whatever
gets shifted out.
As mentioned, TURBO Pascal also provides two functions and a
procedure for byte manipulation:
Lo(I) Returns lower byte of I
Hi(I) Returns upper byte of I
Swap(I) Swaps upper and lower bytes of I
These do just what you think they would do. For example,
assuming that Val = $0810, then
Lo(Val) returns $0010
Hi(Val) returns $0008
Swap(Val) returns $1008
Note that both La and Hi set the upper byte to $00.
9.3 BYTE
The data type Byte is simply a subrange of Integer, consisting of
9-9
the values 0 through 255. Byte and Integer expres'sions and

variables can be freely mixed. In fact, Byte variables act like
Integer variables in all respects but this:any value outside of the
range 0.. 255 which is assigned to a Byte variable is pared down to
the 0.. 255 range. For example, if Small is declared to be of type
Byte, then the following assignments result in the corresponding
values for Small:
Small:= -1; 255
Small:= -2; 254
Small := 256; o
Small:= Round(I040.0); 16
Small:= Round(I04000.o) *error*
In short, assigning a value to a Byte variable is equivalent to
assigning that value to an integer variable, then assigning the
lower byte of that integer to the Byte variable. Confused? Well, if
Large is of type Integer, then Small := -1 is equivalent to:
Large:= -1;
Small := LoCLarge); or Small := Large and $OOFF;
There is one other difference:Byte variables take up only one byte
of memory, while Integer variables take up two. Therefore, if you
want to be economical with your computer's memory and only
need to store small values, you should use Byte variables.
9.4 REAL
Suppose that you want 11/4 to equal 2.75 instead of 2. Suppose

that you want199/200 to equal 0.995 instead of O. Suppose that
you want to find the answer to 502*200. What do you do? You use
real numbers, instead of integers.
Reals differ from integers in two important respects. First, they

have a decimal point with additional digits following it. In other
words, they have values such as 2.75,0.995, -421.0, and so on.
Second, they can (but do not have to) have exponents. The
exponent represents some power of 10 by which the rest of the
real number is multiplied. For example, the value 100400 can be
written as:
9-10
100400.0 or 100400.0E+0 (100400.0 * 1)

10040.0E+1 (10040.00 * 10)
1004.0E+2 (1004.000 * 100)
100AE+3 (100.4000 * 1000)
10.04E+4 (10.04000 * 10000)
1.004E+5 (1.004000 * 100000)
0.1004E+6 (0.100400 * 1000000)
Note the formatmantissa E exponent. The mantissa is simply a
real number, say, 100.4. The exponent is some integer value with
an explicit sign (+ or -), say, +3. And the resulting value is:
<mantissa> * 10 <exponent>
such as 100.4 * 1000, or 100400.0. Of course, you don't always
have to use the exponent format-just if you're working with very
large or very small numbers. (A negative exponent gives you a
very small number, moving the decimal point to the left according
to the value of the exponent.)
If you choose to not use exponents, you can just write real
numbers as you normally would:
3.1415926
-3546.3
0.0034
56793834.21
The same rules of operator precedence mentioned above apply
to real numbers. In addition, TURBO Pascal offers some pre-
defined functions for use with reals. Note that the trigonometric
functions (arctan, cos, sin) assume that the angle involved is in
radians (360 degrees = 2*pi radians):
Abs(x) returns absolute value of x (real)

ArcTan (x) returns arctangent of x (real)
Cos (x) returns cosine of x (real)
Exp(x) returns e to the x power (real)
Frac(x) returns fractional portion of x (real)
Int(x) returns integer portion of x (real)
Ln(x) returns natural log of x (real)
9-11
Random returns random number from 0.0 to 0.99... (real)

Round(x) returns integer nearest to x (integer)
Sin(x) returns sine of x (real)
Sqr(x) returns x*x (real)
Sqrt(x) returns square root of x (real)
Trunc(x) returns integer portion of x (integer)
You need to be careful mixing reals with integer values. TURBO
Pascal will automatically convert integer numbers to reals, but
not vice versa. Let's suppose that you wanted to find the area of a
circle of some radius. You might write this program:
program. FindRadius;
{
note: Pi is a predefined Real constant = 3.1415926536
var
Radius,Area : Integer;
begin
Write('Enter radius' '); ReadLnCRadius);
Area:= Pi * Radius*Radius;
WriteLnC'The area is ',Area)
end; { of program FindRadius }
TURBO Pascal will not compilethis program; instead, you'll get a

"Type mismatch" error. Why? Well, the expression Pi *
Radius*Radius is, by default, a real expression, and you can't
assign a real value to an integer variable. However, you can use
Trunc or Round to make the conversion. Trunc(X) returns the
integer portion of X; in other words, it drops the fractional portion
of X. Here are some sample results:
Trunc(3.1415926) = 3
Trunc(-32.3) = -32
Trunc(421.7) = 421
Trunc(0.5) = 0
Trunc(543832.32) = **error** (too large for integer)
Round(X), on the other hand, returns the integer nearest to X. A
special case occurs when X is exactly between two integers, i.e.,
the fractional portion is 0.5. TURBO Pascal will round away from
9-12
zero, which means Round(22.5) = 23 and Round(-22.5) = -23.

Here are some other results:
Round(3.1415926) = 3
Round(-32.3) = -32
Round(421.7) = 422
Round(O.5) = 1
Round(542832.32) = **error**(too large for integer)
Note that if you use Round or Trunc on a real value that's too
large, you will get a run-time error.
You can now rewrite your program like this:
program FindRadius;
{
note: Pi is a predefined Real constant = 3.1415926536
var
Radius,Area : Integer;
begin
Write('Enter radius: '); ReadLn(Radius);
Area:= Rmmd(Pi*Radius*Radius);
WriteLn('The area is ',Area)
end; { of program FindRadius }
Two issues come up with reals or, rather, two aspects of the same
issue. The issue is this:how many bytes are used to represent a
real? With TURBO Pascal, the answer is:6 bytes. The first aspect
deals with the allowable range for the exponent. Unless you're
dealing with really large numbers, you won't have worry too
much abou~ this-TURBO has a range of -38 to +38, and 10
raised to the power of 38 is a very large number. This, of course,
doesn't mean that the issue will never come up. It's just very
unlikely.
More critical is the issue of precision in the mantissa. In plain

English, how long can the real number be? With TURBO Pascal,
about 11 digits. This provides enough precision to be able to do
financial and other calculations without too much worry. Even so,
you m'ay end up with some slightly (or not-so-slightly) wrong
9-13
values should you push too close to the limit or should you have
to do a long series of calculations. This is due to cumulative
round-off error, or CROE. CROE is a fancy way of saying that bad
values tend to propagate themselves. Too many bad values can
work together to produce results that have no relation to your
initial values.
What causes these mysterious errors? Here are four major
factors behind CROE. First, decimal numbers are represented in
binary form. Did you know that the number 0.1 is impossible to
express exactly in binary form? It's an infinite fraction, much like
1/3 in decimal form (0.333333 ... ). Several of these errors can
cause values to shift slightly.
Second, values with different exponents are normalized before
addition. (This means that their exponents are made equal and
their mantissas adjusted accordingly. If you add 1.0E+04 (10,000)
and 1.5E+OO (1.5), then the latter is normalized to 0.00015E+04
before the two are added together. With limited precision, you
might lose some digits off during' normalization.
Third, when nearly identical values are subtracted, the least
significant (and least accurate) digits become the most significant
digits. If you subtract 1.4356876523E+05 from 1.4356876527E+05,
you should end up with 4.0E-05. But if you have 11 significant
digits, then the last digit in each number (3 or 7) is the one most
likely to be wrong due to the other errors mentioned here. The
result: your answer may be totally wrong. Furthermore, that
answer is then used to "fill" 11 digits. Combine that with the
binary/decimal problem, and you get an actual answer of
3.981590271 OE-05 ... not quite what you'd expect.
Finally, when two numbers of some precision are multiplied

together, the result has twice as many digits, half of which are
then thrown away. For example, if you multiply the two numbers
above, you get 2.06111990355E+10. The "5" at the end represents
the lower 12 digits of the answer (11 digits plus itself). Again, due
to round off and internal representation, errors can creep in,
though this is usually less serious than the pitfalls mentioned
above.
9-14
An error always starts out at the lower end of a number and

moves up towards the front. The more calculations you have, the
farther it can move. The fewer digits of precision you have, the
faster it can move. The moral: if you're using real numbers, be
aware of your limits. If you're going to do a lot of number
crunching, you might want consider using the TURBO-B7 option
(assuming that you're running on an B086-based computer with
an BOB7 math co-processor). This will not only extend your
precision to 16 digits (B bytes), but it also gives you a much wider
exponent range (-307 to +308) and, of course, tremendously
speeds up all your real number calculations.
9.5 CHAR
The data type Char is simply the set of ASCII characters. More
accu rately, a variable of type Char can have 256 val ues, incl ud i ng
the 128 characters in the standard ASCII set. You may remember
from Chapter 7 the discussion on printing and non-printing
characters. Printing characters can be represented as the char-
acter itself within two single quote marks:
'a' '$'
, , 'J'
Remember the control characters? Those are the characters with
ASCII values 0 through 31, and they are usually expressed in
terms of the printing characters ranging from 64 to 95. For
example, the character with the ASCII value of 7 is commonly
called "control-G"; 'G' has an ASCII value of 71 (=7+64). But
since these characters have no printable equivalent, there's no
way to represent them within quotes. However, TURBO Pascal
lets you represent them with the notation:
Aohar,
where char is the corresponding printable character. You can use:

AG for "control-G"
A[ for ESC (ASCII code 27)
and so forth.
9-15
That takes care of codes 0 through 31, but what about characters
with ASCII codes 127 (DEL) through 255? TURBO again comes
to the rescue, with the notation #val, where val is a byte constant
(0 ... 255). For example, you could represent DEL as #127 or #$7F
(yes, you can use hex constants). Here, then, are some different
ways of representing the same characters:
NUL #0 #$00 A@
ctrl-G (BEL) #7 #$07 AG
ESC #27 #$1B A[
blank (space) #32 #$20 ' ,
the digit 0 #48 #$30 '0'
the letter A #65 #$41 'A'
DEL #127 #$7F
ASCII 237 #237 #$ED
Suppose, now, you wanted to put a control-G (which causes a
beep on printing) within a string. If you wrote:
ErrorMsg:= 'ERRDR: AG You entered the wrong value';

then it would print:
ERROR: AG You entered the wrong value
and there would be no beep. Not to worry:TURBO Pascal makes
insertion of special characters simple. All you have to do is to
break the string up into two parts and put the control-G between
them:
ErrorMsg:= 'ERRDR: 'AG' You entered the wrong value';

This automatically inserts the control-G character into the string
contained in Error. You can do the same thing with the #val
characters. In fact, you can build a string consisting of only
special characters:
FiveBeeps:= AQAQAQAQAG#13#lO;
FiveBeeps now has a length of 7 and contains 5 control-G's
(BEL), a carriage return (CR), and a line feed (LF). Writing this to
the screen would cause 5 beeps and would then move the cursor
to the start of the next line.
9-16
TURBO Pascal has a few functions that deal with characters:

Chr(l) returns the character corresponding to I (char)
Ord(C) returns the integer corresponding to C (integer)
Pred(C) returns the character preceding C (char)
Succ(C) returns the character following C (char)
UpCase(C) returns the upper-case value of C (char a... z)
You'll find additional discussions of characters in Chapter 14,
which talks about strings.
9.6 BOOLEAN
George Boole, a 19th-century mathematician, developed a whole

set of rules for numbers that could have only two values:O and 1.lt
was called Boolean algebra. When computers appeared a
century later, most were based on components that could
represent one of two values: 0 and 1. All of Boole's work, and the
work of those who followed him, found immediate application.
Pascal allows you to declare variables of type Boolean. However,
instead of using 0 and 1, Pascal uses False and True to represent
boolean values. Boolean variables and expressions aren't terribly
useful all by themselves; however, they are critical for control
structures, which is the topic of Chapter 10. Boolean expressions
will be covered in more depth there.
Pascal provides the logical operators not, and, or, and xor for use
with boolean values. The first of the two tables at the end of this
chapter shows the precedence for these operators, and the
second table shows what results you get (substituting False for 0
and True for 1).
Boolean values can also be returned by using the comparison
operators to compare other values:
9-17
= <>
<
>
<=
>=
IN
For example
3>5
will return Fa/se, while
7 <= 10
will return True. Again, more about this will be discussed in
following chapters.
OPERATOR PRECEDENCE
Unary minus (arithmetic negation)

not Logical negation
*, I, div, mod, and, shl, shr Multiplying operators
*, -, or, xor Adding operators
=, < >, <, >, <= , >=, in Relational operators
LOGICAL OPERATOR TRUTH TABLES
o= False 1 = True
exp1 exp2 (exp1 and exp2) (exp1 or exp2) (exp1 xor exp2)
0 0 0 0 0
0 1 0 1 1
1 0 0 1 1
1 1 1 1 0
9-18
Part 2 Chapter 10 Control Structures
10. CONTROL STRUCTURES
One of the great buzz words (buzz phrases?) of computer

science is flow of control. This term refers to the order in which
the statements in your program are executed. Up until now,
you've only looked at sequential flow of control. Simply put,
sequential flow of control means that each statement is executed
exactly once, starting with the first statement and ending with the
last. Nice, but it won't get you far.
To provide a little variety, Pascal has a number of control
structures, special constructs that allow you to have nonsequen-
tial execution of statements. There are three types of control
structures in Pascal: conditional, iterative, and case.
The goal of this section is to explain what these constructs are
and how you use them. First, though, you need to look at a few
basics: statements and boolean expressions.
10.1 STATEMENTS·
The primary unit of execution in a Pascal program is the

statement. At the simplest level, there are two kinds of statements:
procedure calls and assignments. You've looked at built-in
procedures, such as WriteLn; a procedure call is the act of using
such a procedure:
WriteLnC'The sum is ',C);
Assignment statements assign some value to a variable. They
take the form:
<variable> := <expression>;
where expression is, well, an expression that resolves to a value
appropriate to the data type of variable. For example, if the
variables Score and Maximum are of type Integer, then the
following statements are valid:
10-1
Chapter 10 Control Structures Part 2
Score:= 10;
Maximum := 32*Score div 17;
Score := Score + 10;
Maximum := Succ(Maximum);
All but one of the control structures we'll be looking at will

execute only one statement. This doesn't sound very impressive
or convenient, and, in fact, it would make for a very serious
limitation if you could only use procedure calls and assignment
statements. Such is not the case. Pascal considers the following
structure to be a single statement:
begin
<statement 1>;
<statement 2>;
<statement n>
end;
This structure is known as a block or compound statement. It can

be used anywhere that a simple statement can; you can even use
one block statement within another. Compound statements are
most often used in conjunction with control structures, as you'll
soon see.
-Blocks (compound structuresJ-
10-2
10.2 BOOLEAN EXPRESSIONS

A moment ago, you were looking at expressions-integer
expressions, in particular. Thoseshown were integerexpressions
because they returned integer values. A boolean expression,
then, returns a boolean value-in other words, either True or
False. For example, suppose that the integer variables Score and
Maximum have each been set to some val ue, say, Score = 10 and
Maximum = O. You can now create the following boolean
expressions, with the corresponding results:
Score >Maximum True (is greater than)
Score =Maximum False (is equal to)
Score <Maximum False (is less than)
Score >= Maximum True (is greater than or equal to)
Score <= Maximum False (is less than or equal to)
Score <> Maximum True (is not equal to)
In a Pascal program, such expressions would actually return
True or False, the two possible values of the predefi ned data type
Boolean. We could define a boolean variable, NewMax, and then
have the following assignment statement:
NewMax:= (Score> Maximum);
Depending upon the current values of Score and Maximum, this

would set NewMax to either True or False. For example, given the
values defined above, NewMax would receive a value of True.
You can create more complex expressions by using the Boolean
operators and, XOR, or, and not. not flips the value, changing
True to False and vice versa. and, xor and or combine two
expressions to create a new one. The table at the end of this
chapter is a truth table showing how these operators work.
Using these operators, you can create such expressions as
(Score> Maximum) or (Score> 30000)
(Score> 10000) and (Score <= 20000)
not (NewMax or (Maximum = 0))
10-3
If we assume that Score, Maximum, and NewMax have the values

given above (10, 0, and True), then the first expression resolves to
True, while the other two resolve to False. Having discussed all
this, you're now ready to go on to control structures. {Decisions
Illustration goes here)
-Decisions, Decisions, Declslons-
10.3 CONDITIONAL EXECUTION
In writing programs, you usually have groups of statements that

you want to be executed only if some condition has been met.
Not surprisingly, this is known as conditional execution. In
Pascal, conditional execution takes the form:
if <boolean expression>
then <statement>;
10-4
The if. .• then structure works quite simply. If boolean expression

returns a value of True, then statement is executed; otherwise, it
is ignored. For example,
if Score> Maximum
then Maximum := Score;
will set Maximum to Score if and only if Score is greater than
Maximum. You could also do the following:
NewMax := (Score> Maximum);
if NewMax
then Maximum := Score;
If you really want to get fancy, you can do this:
if Score> Maximum then begin
Maximum := Score;
WriteLn('Congratulationsl' );
WriteLn('Your new high score is ',Maximum)
end;
Tied closely to the idea of conditional execution is that of

alternation. Alternation means that one set of statements is
executed if a condition is True, and another set ifthat condition is
False. You could do it this way:
if <condition>
then <statementl>;
:If not <condition>
then <statement2>;
Pascal, however, provides a simpler way:
:If <condition>
then <statementl>
else <statement2>j
By using the if...then ... else construct, you can amplify your
example as follows:
10-5
if Score> Maximum then begin

Maximum := Score;
WriteLnC 'Congratulations!');
WriteLnC'You have set a new high score of ',Maximum)
end
else begin
WriteLnC'Your score was ',Score);
WriteLnC'Nice try, but you can do better')
end;
10.4 ITERATION
Conditional execution adds a lot to your programming capabili-

ties, but you still have a major handicap: You can execute each
statement only once. You need the ability to repeat a section of
code some number of times or until some condition is met. This
looping process is known as iteration. Niklaus Wirth must have
really liked loops, because Pascal has not one, not two, but three
forms of iteration: for... do, while ... do, and repeat. .. until. Let's look
at each.
-Looping-
10-6
10.4.1 The FOR ... DO Loop
There are situations where you want to have a statement

executed some number of times. Often, you want or need to
know how many times you've gone through the loop so far. Or,
you might want a variable to step through a range of values,
executing a statement (or group of statements) once for each
value.
The for... do loop allows you to do this. It uses the format:
for <var> := <expl> to <exp2> do
<statement>;
When the loop is encountered, var is assigned the value ofexp1. If
var is less than or equal to exp2, then statement is executed. var is
then increased by one, and the test is repeated. This goes on until
var is greater than exp2. If exp1 is greater than exp2, then
statement is never executed. For example, you could write:
for Indx:= 1 to 10 do
WriteLnC'n = ' ,Indx,' n*n = ' ,Indx*Indx);
This will display the integers from 1 to 10, along with their
squares. If you want to decrement (decrease) var instead of
incrementing it, you can use the format:
for <var> := <expl> downto <exp2> do

<statement>;
Now, var decreases each time until it is less than exp2, and
statement is never executed if exp1 is less than exp2.
The for... do loop in Pascal sacrifices convenience for a gain in

power. You cannot specify a step value by which to change var,
as you can in most other languages. To do that, you have to
create your own loop using one of the other loop constructs
(described below). Also, var cannot be of type Real; again, you
have to use another type of loop.
So much for the inconvenience; where's the power? Well, var
doesn't have to be of type Integer; it can be of any scalar data
10-7
type: Char, Byte, Boolean, or any declared scalar type (DST) you
care to define. Given appropriate definitions, the following loops
are all valid:
for Indx := 30 downto 20 do
<statement>;
for Ch := 'A! to 'Z' do
<statement>;
for Flag := True downto False do
<statement>;
for D~:= Mon to Fri do
<statement>j
Now you can see why there is no step capability. Mixing numeric
and non-numeric values could be very confusing, especially for
those programmers who write compilers. Besides, there is
something clumsy about a statement like
for Month:= January to December step 2 do
<statement>j
while the statement
for Month := January to December step February do
<statement>j
doesn't make any sense at all. So there is no step capability in for
loops.
As with if ... then ... else, you can use a compound statement in
place of statement, to wit:
for Indx := 20 to 30 do begin
<statement 1>j
<statement n>
end;
10-8
10.4.2 The WHILE ... DD Loop
Since the for... do loop can't easily meet all your iterative needs,
Pascal give you two others, both controlled by boolean expres-
sions. The first is the while... do loop:
while <boolean expression> do
<statement>;
This loop will cause statement to be executed as long as boolean
expression resolves to a value of True. It will never be executed at
all if boolean expression is initially False.
If you wish, you can replace for... do loops as follows:
<var> := <exp1>;
while <var> <= <exp2> do begin
<statement 1>;
<var> := suce(<var> )
end;
If you want to use a step value, then you need to change the last
statement to reflect that. For example, to have a step value of 3,
you could change the last statement in the loop to read
<var> := <var> + 3
Of course, you can do other things than imitate for loops. For
example, here's a binary search program. It searches a sorted list
for a given value until it either finds it or determines that it is not in
the list. The advantage of a binary search is that it only takes a
maximum of (n log 2) tries to search a list of n objects. This means
that a list of 1024 elements will take at most 10 tries through the
loop. .
:procedure BinarySearch(DKey: ListVal;

var Index : Integer; var Found: Boolean);
{
This procedure does a binary search for DKey in List,
where List = ARRAY[IMin..IMax] OF ListVal
List is sorted such that List[I] < = List[I +1]
It searches for Index such that DKey = List[Index]
If fmmd, then Found = True else Found = False
10-9
var
Upper,Lower : Integer;
begin
Upper := IMa.x;Lower :=. IMin; Found := False;
while DOt Found and (Lower <= Upper) do begin
Index:= (Upper +;Lower) DIV 2;
if DKey < LiBt[Index]
then Upper :=Pred(Index)
elle ifDKey> LiBt[Index]
.thaD. Lower := Succ(Index)
elle Found:= True
8Dd.
end; { of proc Bina.rySearch }
10.4.3 The REPEA T... UNTIL Loop
The third loop is repeat... until. It takes the format:
repeat'
<statement 1>;
<statement 2>;
<statement n> .
until <boolean. condition>;
This differs from while..• doin three important ways. First, the
statemef)ts within the loop are always executed at least once. In
other words, ,all the statements in the loop are executed before
boolean condition is resolved. Second, the statements continue
to execute as long as boolean condition is false; the while loop
executes while boolean expression is true. Third, the repeat loop
can directly execute more than one statement and, therefore,
doesn't n(3ed a compound statement (begin...end) to handle
multiple statements ..
,One common use of 'repeat loops is to condition input, that is, to
prompt the u~er for some value and to continue to prompt until
the value entered is one of those that you decide to allow. For
example, the following routine prompts the user to enter an
10-10
integer within the range [Iow ... high]. It repeats the prompt if (1)
there is an I/O error (due to bad numeric format), or (2) the value
is not within the proper range:
program TestGetInteger;
type
Prompt = striDg[80];
procedure GetInteger(var Val: Integer; Mag: Prompt;

Low,High : Integer);
begin
{$I-,R-} { turn off I/O, range checking}
repeat
Write(Mag);
ReadLn(Val)
untU (IOresult=O) aDd. (Low<=Val) aDd. (Val<=Htgh)
{$I + ,R+} { turn everything back on }
end; { of proc GetInteger }
This provides "bulletproof" integer input. No matter what the

user tries to do, this routine will continue to ask for a proper value '
until the one entered has the proper format and is within the
proper range.
10.5 THE CASE STATEMENT
When writing programs, you sometimes want to perform one of a

number of actions depending upon the curr~nt value of a
variable. For example, you might want to print out a menu, accept
the user's choice, and then act accordingly. One way you can
handle this is to use repeated if... then ... else st~terr~nts: '
10-11
repeat
Write(,Enter direction: U)p,D)own,L)eft,R)1ght:');
ReadLnCOh) {Oh is of type Ohar}
untU COh IN ['U' ,'U: ,'D' ,'d.' ,'L' ,'1' ,'R' ,'r' J);
if COh = 'U') or COh = 'U:)
thenY:=Y-l
else if COh = 'D') or COh = 'd')
thenY:=Y+ 1
else if COh = 'L') or COh = '1')
then X :=X-l
else ifCOh = 'R') or COh = 'r')
then X := X + 1;
GotoXYCX,Y)
However, the if... then ... else chain can get a little tiresome (and
difficu It to follow) if there are a lot of different conditions to check.
So Pascal includes an additional flow-of-control structure called
the case statement.
The case statement requires a scalar variable which has been set
to some value. Scalar variables include those of type Integer,
Byte, Char, Boolean, and any declared scalar type (DST).
(Variables of type Real are not considered scalar.)The case
statement has this format:
case <variable> of
<constant list 1> : <statement 1>;
<constant list 2> : <statement 2>;
<constant list n> : <statement n>

end;
where a constant list is a list, separated by commas, of constants

of the same data type as the variable. When the case statement is
encountered, the program compares the current value of the
variable against each of the constant lists. When a match is found,
the corresponding statement is executed and the rest of the case
statement is skipped. You could now rewrite your first example as
follows:
10-12
repeat
Write('Enter clirection; U)p,D)own,L)eft,R)ight:');
ReadLn(Ch) {Ch is of type Char}
untU ( Ch IN ['0' ,'U: ,'D' " d' ,'L' ,'1' ,'R' ,'r' j);
caseChof
'U','U: : Y:=Y-l;
'D','d': Y:=Y+ 1;
'L','I' : X:=X-l;
'R','r' : X:=X+ 1
end;
The above example ensures that Ch would only have one of the
specified values in the case statement. This is done by checking
Ch against a set of allowable values (you'll learn about sets in a
later chapter). What if you removed that restriction? What if Ch
had a value other than those listed in the case statement? Under
the original Pascal definition (by Wirth and Jensen), the result
was undefined; that is, the outcome was uncertain.
TURBO Pascal, however, takes the commonsense approach
that, in such a case, all of the statements are skipped and
execution resumes right after the end of the entire case statement.
In fact, TURBO Pascal goes one step farther and allow an else
section at the end of the list of choices. This section is executed
only if the variable does not match any of the constant lists. With
such an extension, you might rewrite your example as follows:
Write('Ent.er clirection; U)p,D)own,L)eft,R)ight: ');
ReadLn( Ch);
caseChof
'U','U: : Y:=Y+l;
'D','d' : Y:=Y-l;
'L','l' : X:=X-l;
'R','r' : X:=X+ 1;
elae
WriteLn('Illegal entry ');
WriteLn('PIease use U ,D,L or R')
end;
10-13
Chapter 10-Control Structures~ Part 2
", . ~\' r . ' : '.
. ,
1Q.6 SAMPLE PROGRAM

Now that you~ve" looked at all of the different flow-of-control
structures, let's put them all together in a program. The listing at
the end of this chapter shows a simple (very simple!) Pascal
program that uses all of the constructs you've learned about.
The'first construct you encounter is a while loop. It tests the
boolean variable Done and will loop as long as Done has a value
of False. . ,
The next construct is a repeat lo~p. The statements in this loop

prompt the u'ser for acharacter. The program will not leave the
loop until the user enters 'one 'of the valid command characters.
"
Withinths'repeat loop is your third construct, an if... then ...else

~tatemeht that tells the userwhether or not one of the command
characters has been ent,ered.
Following the· repeat, loop is your fourth construct, a case
statement. It' performs one of six actions depending upon the
character value the user has entered. Four of the commands
change the values of Xor Y. One command sets Done equal to
True and so allows the program to exit the while loop.
T,heremaining command ('F' or 'f') executes a block statement

which contains your last construct, a for loop. This loop executes
the statement Writer Dot, Bell) five ti mes (the val ue of the constant
FireMax).
10.7 CONCLUSION
These control structures greatly add to your programming

capabilities. They help you to escape from the "every statement
exactly once" approach of your earlier programs while main-
taining a clearly-defined path of execution through the code. By
combining the various structures together, you can create
exactly the path you need.
10-14
Even with these aids, you still are limited to writing one big
program and having to duplicate code if you need it in more than
one place. In the next chapter, you will learn about the solution:
procedures and functions. You've already seen some standard
Pascal procedures and functions, such as WriteLn and ReadLn;
in the next chapter, you'll learn. how to write your own.
Before we go on to the next chapter, here are the table and the
sample program I promised you a while ago:
BOOLEAN TRUTH TABLES
Exp1 Exp2 (Exp1 and Exp2) (Exp1 xor Exp2) (Exp1 or Exp2)
False False False False False
True False False True True
False True False True True
True True True False True
CONTROL STRUCTURES PROGRAM EXAMPLE
program. Example;
{
Example Sample Pascal program
demonstrates control structures
note the use of typed constants
}
canst
Bell =AGj
FireMax = 6;
Dot =" . ,.
Done : Boolean = False;
X : Integer = 0;
y : Integer = 0;
CmdSet : set of Char = ['L' ,'R' ,'U' ,'D' ,'F' ,'Q' ] j
var
Ch : Charj
Count : Integer;
10-15
begin
while DOt Done do begin
WriteLn('X,Y = (' ;X,',' ,Y,'),);
repeat
Write('CMD: L)eft, R)ight, U)p, D)own, F)ire, Q)uit: ');
Read(Kbd,Ch); WriteLn;
if Ch >= 'a' { convert to upper case }
thenCh:= Chr(Ord(Ch)-32);
if DOt (Ch IN CmdSet)
then WriteLn('Illegal entry-try again')
else WriteLn('Command accepted')
untU Ch in CmdSet;
caseChof
'U' Y:=Y+ 1;
'D' Y:=Y-1;
'R' X:=X+ 1;
'L' X:=X-1;
'F' begin
Write( 'Firing');
for Count := 1 to FireMax do
Write(Dot,Bell);
WriteLn
end
'Q' : Done:= True
end { of CaBe }
end { ofwhlle }
end. { of program Example}
10-16
Part 2 Chapter 11 Procedures and Functions
11. PROCEDURES AND FUNCTIONS
In previous chapters, you learned how to write simple Pascal

programs. Period. True, you've learned all about boolean ex-
ressions, flow-of-control structures, and other such esoterica.
But the fact remains that all you can do is write one big program
that starts at begin and finishes at end. If you have a set of
commands that you want to use in five different places, then you
would have to write that code in all five places. (Of course, if you
merely want to execute the same code five times, you could put in
a for loop, but we covered that in the previous chapter.)
That limitation ends now. This chapter will discuss the use of
procedures and functions, which together are known as sub-
programs. You'll first hear about the general concept of a
subprogram. You'll then see how (and where) to place sub-
programs in your main program. Then you'll touch on the ideaof
scope, which governs where a subprogram may be called from.
You'll end with a discussion on parameters for subprograms and
I'll give you a taste of some ofthe different things you can do with
them.
11.1 SUBPROGRAMS
As you've learned in previous chapters, a Pascal program has the

following structure:
program Whatever;
<declaration section>
begin
<program statements>
end. { of program whatever}
As mentioned above, you can get by with just this for small
programs, but you'll find this structure very limiting with big
programs, especially if you have sections of code that are
repeated throughout the program. The solution? Pascal allows
you to create subprograms. A subprogram is indeed like a
11-1
Chapter 11 Procedures and Functions Part 2
miniature program contained within a larger one. Here, for

example, is the structure for defining a procedure (one of two
types of subprograms):
procedure Whichever;
<declaration section>
begin
<procedure statements>
eDd; { of procedure Whichever}
Look familiar? The only visible differences between this format
and the format for a program are (1) the word procedure is used,
and (2) the end is followed by a semicolon instead of a period. (A
few more differences will crop up later.) As with a program,
execution starts right after the begin statement and fi nishes at the
end statement.
The next question is: where do you put subprograms in your
program? Answer: after the declaration section and before the
opening begin of your program. Indeed, you can think of your
subprograms as being the last part of your declaration section
(after all of your label, const, type," and var statements). Some of
you may then wonder if you can declare a subprogram within a
subprogram. Yes, you can, just like you would ina regular
program. And you could declare subprograms within that
subprogram, and so on ad infinitum (well, not quite; the TURBO
Pascal compiler does have its limits).
Now that you've got the subprogram in place, the next question
is, how do you use it? Answer: simply by stating its name. Let's
put our procedure within our program and call it a few times:
program whatever;
cODSt
Maximum = 10;
type
ORange = O..Max1mum;
var
Indx : ORange;
11-2
procedure Whichever;
cODBt
Star = '*",
type
DIW8 = (Mon,Tues,Wed,Thur,Fri,Sat,8un);
var
Daor : DlW8i
begin { main boqy of proc Whichever}
for Daor:= Mon to Fri do
Write(Star)
end; { of proc Whichever}
begin { main boqy of program Whatever}

for Indx := 0 to Maximum do begin
Whichever;
WriteLn
end
end. { of program Whatever}
In this example, const, type, and var statements have been

deliberately included in both the main program and the sub-
program so that (1) you won't be confused about where
subprograms go, and (2) you can see how similar the two
structu res are. Look over the example carefu lIy and see if you can
figure out just what it will do. In case you're not sure, remember
that Write and WriteLn are built-in procedures (Le., they're always
there-you don't have to definethem). Write(Star) will print an '*'
on the screen, while WriteLn will start a new line.
Finished? If you guessed that this program would print out eleven
lines, each consisting of five asterisks ('*****'), you're right. The
procedure Whichever prints out five asterisks each time it is
called. The program Whatever calls Whichever eleven times,
starting a new line after each call.
11.2 SCOPE
Before you can go any farther into the intricacies of sub-

programs, you must first deal with the concept of scope. In a
BASIC program, all identifiers (variables, functions, line numbers)
are global in scope. In other words, any identifier can be
11-3
referenced from anywhere in the program. Indeed, any identifier

can be declared anywhere in the program, and variables are
declared merely by referring to them.
In Pascal, things are quite different. The Great Underlying Rule of
Pascal is that a given identifier must be declared before it is used.
Again: any constant, data type, variable, or subprogram you care
to use has to be explicitly defined first.One last time:each and
every identifier appearing in any program (or subprogram)
statements needs to have been previously described in a declar-
ation section.
Having made that pronouncement, you must now understand

what is meant by "previously declared." In the simplest sense, it
means that if you start at the program statement and move down,
you must run across the declaration of each identifier before you
find any other use of it. You can easily check that in a program
containing no subprograms, since you need only to verify that
each identifier used in the program statements is found in the
declaration section (and that, of course, you are following the
same rule within the declaration section itself). But consider this
example:
program Scopel;
var
A Integer;
procedure SetB;
var
B Integer;
begin { proc SetB}
B:=2*A
end; { ofproc SetB}
begin { main body of prog scopel }

A:=2;
SetB;
WriteLn(A+B)
end. { of program Scopel }
11-4
Here, your simple rule has been met-both A and B have been
defined as being variables of type Integer before being used-
and yet this program will not compile. Why? Because B, having
been declared in the procedure SetB, is only recognized within
that procedure. In other words, the variable B is local to the
procedure SetB. By contrast, the variable A is recognized both
within SetB and in the main program itself.
So you now must modify your simple rule to this:each identifier
must have been previously declared in an enclosing program or
subprogram before being used. By enclosing, I mean that .the
definition of the identifier must come after its declaration and
before the begin statement of the program or subprogram in
which is was defined. In the example above, you can see that the
statement WriteLn(A+B) comes after the begin statement of the
procedure SetB.
Here's another case to consider:
:program Scope2;
var
A Integer;
:procedure SetA;
var
A Integer;
begin
A:=4
end { of proc SetA }
begin { main body of program Scope2 }

. A:=3;
SetA;
WriteLn(A)
end. { of prog Scope2 }
Two questions: (1) will this program compile, and (2) if it does,
what value will it print out? Two answers: (1) yes, and (2) 3. This
brings up a corollary to our rule:the most recent declaration of a
given identifier will always be used. The procedure SetA assigned
its own local variable A to 4 and left the global variable A
untouched.
11-5
An editorial comment. Those of you used to BASIC or FORTRAN

may chaff at this discipline that Pascal requires of you. After all,
it's so nice to make up variables as you go along. However, an
excellent reason for these restrictions exists. There are two great
sources of program bugs in BASIC and FORTRAN: misspelled
variables and side effects. It is not uncommon to type "1 NDEX" in
one statement and mis-type "INDX" in another and have the
program (BASIC or FORTRAN) g6 merrily on its way. It is also
not unusual to (unknowingly) use the same variable in several
different parts of the program, often with bizarre results
(especially if those sections call one another). Pascal takes great
strides towards eliminating those bugs. You can still create them,
it is true, but you will really have to work at it.
11.3 PARAMETERS
Perhaps the most common use of subprograms is to perform the

same set of operations using different sets of values. For
example, you might want a procedure which would swap two
integer values. The question is, which two values should it swap?
One solution is this:
program. TestSwap;
var
Sl,S2,Alpha,Bravo,Charlie,Delta,Eagle,Foxtrot
: Integer;
procedure Swap;
var
Temp Integer;
begin
Temp:=Sl;
Sl:= S2;
S2 :=Temp
end; { ofproc Swap}
begin { ma1n body of program TestSwap }

Alpha:= 1; Bravo:= 2; Charlie:= 3;
Delta:= 4; Eagle:= 5; Foxtrot:= 6;
11-6
81 := Alpha; 82:= Eagle;

8wap;
Alpha := 81; Eagle := 82;
81 := Bravo; 82 := Delta;
8wap;
Bravo:= 81; Delta:= 82;
81 := Charlie; 82:= Foxtrot; .
8wap;
Charlie := 81; Foxtrot := 82
end. { of prog 8wapTest }
Now, a little thought will show just what a lousy approach this is.
You are executing more statements before and after each call to
Swap then you would if you just did each swap directly. Luckily,
you don't really have to do this. Instead, you can define Swap as
having a list of parameters. For example:
program TestBwap;
var
Alpha,Bravo,Charlie,Delta,Eagle,Foxtrot
: Integer;
procedure 8wap(var 81,82: Integer);

var
Temp : Integer;
begin { proc 8wap }
Temp:=81;
81:= 82;
82 :=Temp
end; { ofproc 8wap}
begin { main body of program TestBwap }

Alpha:= 1; Bravo:= 2; Charlie:= 3;
Delta := 4; Eagle:= 6; Foxtrot := 6;
8wap(Alpha,Eagle);
8wap(Bravo,Delta);
8wap( Charlie,Foxtrot)
end. { ofprog TestBwap }
A lot better, no? The integer variables S1 and S2 are no longer

global variables but instead are parameters of the procedure
11-7
Swap. Each time Swap is called, the actual variables in the call
(say, Alpha and Eagle) are used in place of S1 and S2. Because of
this, S1 and S2 are called formal or dummy parameters, while
Alpha and Eagle are known as actual parameters.
There is one other twist in parameter lists. The prefix var before
any formal parameter means that the corresponding actual
parameter is directly substituted for the formal parameter. In
other words, if the formal parameter has its value changed, the
actual parameter is also changed. If there is no var prefix, then the
actual parameter cannot be affected by anything that happens to
the formal parameter. For example, here's a procedure that
forces a variable to be within two values:
procedure Condition(Min : Integer; var Valu : Integer;
Max: Integer);
begin
it Min > Max
then SwapcMin,Max);
itValu<Min
then Valu:= Min
else itValu > Max
then Valu := Max
end; { of proc Condition}
Why do you want to have formal parameters that do not return

values? Well, if you make the call Condition(20,Alpha,O), you
don't want the procedure to return a value to the constants "20"
and "0", and, indeed, Pascal would not let you. If you defined
Condition with the parameter list (var Min, Valu,Max : integer),
then you would get a compiler error with the above call. Why?
Because the compiler knows that it cannot return a value to a
constant or an expression, but only to a variable.
11.4 FUNCTIONS
As you've seen, a procedure can change the values of parameters
passed to it. The calling program (or subprogram) can then use
those modified parameters for whatever purpose it has. Suppose,
forexample, that you wrote a procedure to find the integersquare
root of a given value:
11-8
:procedure ISqrt(Valu: Integer; var Root: Integer);

var
OddSeq,Sqriare : Integer;
beg:in
OddSeq:= -1;
Square:=O;
repeat
OddSeq:= OddSeq + 2;
Square := Square + OddSeq
untU Valu < Square;
Root := Succ(OddSeq div 2);
ifValu <= Square - Root
then Root := Pred(Root)
end; { of proc ISqrt }
This procedure would take Valu, find its square root, and set Root
to that value. The calling program might use it as follows:
re:peat
Write('Enter value: '); ReadLn(Able);
ISqrt{Able,Baker);
WriteLn( 'The square root is ',Baker)
untUAble = 0;
There are many other cases where you might want to do
something similar, that is, return a value based on certain
variables or conditions. After a while, it can get a little clumsy
and/or tiring to always set aside one parameter for the returned
value. So, instead, you can use a function. A function acts just like
a procedure, but with one difference: it returns a Real or a scalar
value. (Brief review:scalar data types include I nteger, Byte, Char,
Boolean, and any defined scalar type-OST.) For example, you
might rewrite ISqrt as follows:
function ISqrt{Valu: Integer): Integer;
var
oddSeq,Square,Root : Integer;
beg:in
ISqrt := Root
end; { of func ISqrt }
You could now use it in your program like this:
11-9
repeat
Write(/Enter value: '); ReadLnCAble);
WriteLn(/The square root is I,ISqrt(Able))
until Able = 0;
A function can be used anywhere that a constant or an
expression of the same data type could be used. Suppose you
wanted to find both the integer square- and fourth-roots of a
given value. You could rewrite the program this way:
repeat
Write(/Enter value: '); ReadLnCAble);
WriteLn('The square root is ',ISqrt(Able));
WriteLnC/The fourth root is I,ISqrt(ISqrt(Able)))
until Able = 0;
When using a function, care must be taken to ensure that you
have set it to some value before exiting. As shown above, you do
this by assigni ng some value to the function name. You must also
be careful about using the function name in any other way within
the function itself. For example, you couldn't rewrite ISqrtto look
like this:
procedure ISqrt(Valu: Integer; var Root: Integer);
var
oddS eq,S quare : Integer;
begin
OddSeq:= -1;
Square :=0;
repeat
OddSeq:= OddSeq + 2;
Square := Square + OddSeq
until Valu < Square;
ISqrt:= Succ(OddSeq div 2);
ifValu <= Square - ISqrt
tli.en ISqrt := Pred(ISqrt)
end; { of proc ISqrt }
This would cause a compiler error. Why? Because Pascal

considers the other uses of ISqrt in this statement to be additional
calls to the function itself, and you don't have a parameter for
each one. "Can a function call itself?", you might ask. Yes, it can ...
11-10
11.5 RECURSIVE SUBPROGRAMS
As mentioned above, the Great Underlying Rule of Pascal is:no

identifier (constant, data type, variable, or subprogram) may be
used unless it has been previously declared in the main program
or in an enclosing subprogram. The corollary is that all such
identifiers may be used, which means that a subprogram can call
itself. For example, let's suppose you're writing a graphics
package and want a procedure which will fill a blank area on the
screen. Let's also suppose that you've already written two other
graphics subprograms: PlotXY(X, Y), which plots a point at X,V;
and Plotted(X, Y), a boolean function which returns True if the
point X,V has been plotted, and False otherwise. Here, then, is
your procedure:
procedure Fill(X,Y : Integer);
begin
if DOt Plotted(X,Y) then begin
PlotXY(X,Y); { plot this location}
Fill(X+l,Y); { check point to the right}
Fill(X-l,Y); { check point to the left }
Fill(X,Y+1); { check point below}
Fill(X,Y-l) { check point above}
end
end; { of proc Fill }
Can you see how this works? Each time Fill is called, it checks to
see if the location (X,V) has been plotted. If it has, then Fill doesn't
do anything. Otherwise, Fill plots a point at (X ,V) and then tries to
fill in each of the four adjacent points to (X,V).
One cautionary note about recursive subprograms: they can
quickly use up a lot of memory, causing your program to behave
erratic'ally or to blow up. You could easily come up with a
example where Fill would continue to call itself until you had over
100 levels of nested subroutine calls ... and a lot of systems just
can't handle that many levels,of nesting. So be careful.
11-11
11.6 FORWARD DECLARATIONS
Are there any more situations where the Great Underlying Rule of
Pascal gets in the way? Well, occasionally, you may wish to call a
subprogram before it has been declared, out of convenience or
necessity. For example, you might have the following program:
program Example;
var
Alpha : Integer;
procedure Testl(var A: Integer);
begin
A:=A-l;
itA>O
then Test.2(A)
end; { of proc Testl }
procedure Test.2(var A: Integer);

begin
A:=ADIV2;
itA>O
then Testl(A)
end; { of proc Test.2 }
begin { main body of program example}

Alpha:= 15;
Testl(Alpha)
end.. { of proc Example}
As you can see, Test1 calls Test2, and Test2 calls Test1. As it
stands, this program won't compile; you'll get an "Unknown
identifier" error when it finds the reference to Test2 within Test1.
If you swapped Test1 and Test2, then you'd get a similar error
within Test2. So vyhich do you declare first? The answer:both.
How? By declaring Test2 with a forward declarations:
program Example;
var
Alpha : Integer;
procedure Test.2( var A : Integer); forward;
11-12
:procedure Test1(v8l'A: Integer);

beg:ln
end; { of proc Test1 }
»rocedure Test2;
beg:ln
end; { of proc Test2 }
beg:ln { main boqy of program example}

Alpha:= 16;
Test1(Alpha)
end. { of proc Example}
The forward declaration of Test2 contains only that information

necessary to resolve any references to it, namely its name and
parameter list. The actual body of Test2 (which, you will notice,
no longer has a parameter list) occurs after Test1. Now, Test1 can
call Test2 (because of the forward statement) and Test2 can call
Test1 (since the latter precedes the former).
11.7 EXTERNAL SUBPROGRAMS
TURBO Pascal allows you to call assembly language procedures

and functions that you have assembled separately. These are
known as external subprograms; you can learn more about them
in Chapter 22.
11-13
Part 2 Chapter 12 Defined Scalar Types
12. DECLARED SCALAR TYPES
In Chapter 9, we talked about four scalar data types: Integer,

Byte, Boolean, and Char (Real isn't really a true scalar data type).
As you may recall, scalar simply means that each data type has a
fixed range of distinct values; for example, the data type Byte has
256 different values (0 ... 255), while Boolean has exactly two
(False and True). These four data types are very useful, but
sometimes they aren't quite enough. Let's look at a brief example
to see why:
program Example;
var
Month,Daor;DB80fWeek,Year Integer;
begin
Year:= 1989;
Month:=8;
Daor:= 24;
DB80fWeek := 3;
These values together represent some date. The year is obviously

1989, and the day is the 24th of some month-but what is the
month, and what day ofthe week is it? You can assume (probably
correctly) that the programmer is using standard numbering for
the months (1=January, 2=February, etc.) and by some quick
counting come up with "August". But the "day of the week"
question is even harder to deal with. What day does the
programmer assume is the first? Is that day numbered "0" or
"1"? Unless you know the answer to those two questions, the
statement "DayOfWeek := 3" doesn't really tell you anything.
This is a common problem in programming. You use a range of
numbers-suc'h as 1 through 12-to represent some corres-
ponding set of values-such as January through December. But
deep into the program, you lose track of what numbers represent
what values. For example, the value you assigned DayOfWeek
12-1
Chapter 12 Defined Scalar Types Part 2
could easily represent Tuesday, Wednesday, or Thursday, de-

pending upon whether you started the week on Sunday or
Monday, and whether you used a or 1 to represent that first day.
As it turns out, August 24th, 1989, falls on a Thursday-so your
example assumes that Monday is the first day of the week and
equals O.
Now consider the following variation of the example:
program Example;
type
Months = (January,February,March,April,Maor,June,July,
August,8eptember,October,November,December);
Weekda8s = (Monda8,Tuesct.aor,Wednesda8,Thursda8,Frlda8,
Saturct.aor,8unct.aor);
var
Year,Daor Integer;
Month Months;
DBiYOfWeek Weekct.aors;
begin
Year := 1989;
Month := August;
Daor:= 24;
DBiYOfWeek := Thursd.a¥;
Now the code is completely clear as to the date and day of the
week. All uncertainty and ambiguity is gone. Instead of having to
second-guess the programmer, you can easily see what each
statement means.
In defining Months and Weekdays, you have created two
declared scalar types (DSTs). To make a DST, you simply (1)
come up with a name for it, and (2) enter its list of values. The
name and the values both follow the rules for creating identifiers:
starts with a letter or underscore, followed by 0 or more letters,
digits, and underscores; up to 127 characters long; case (upper-
and lower-) doesn't matter; all characters (including under-
scores) are significant. The DST is defined in a type section; it has
the format
12-2
<name> =«first value>, ... ,<last value»;

The DSTs defined in the example were nice but not completely
necessary. Since both months and days of the week are
sequential lists, translation is easy once you know where to start.
If you know that Monday is the first day of the week, and that its
corresponding value is 0, then it's simple to equate "3" with
"Thursday". But sometimes you work with arbitrary lists of
values, values which don't really fall into an apparent order. Let's
look at a second example:
program Example2;
var
ShipClass Integer;
begin
ShipClass := 7;
end.
Assuming that this is a portion of a "Trek" program, then

Shipe/ass probably tells you something about what type of
starship you are dealing with. The question is, what type are you
dealing with? Unlike months and days of the week, there is no
one way of ordering a list of different classes of starships. You
can't even make a guess; you just don't have enough information.
Now, let's look at the same example using a DST:
program Example2;
type
ShipType = (NoShip,Constitution,Enterpr1se,Reltant,
Loknar,Larson,Chandley,Excelsior)3aker);
var
ShipClass : ShipType;
begin
ShipClass := Reliant;
12-3
Once again, you now have no question as to what type of ship

you're working with. Careful, clear definition of DSTs can greatly
improve the readability of the programs you write.
Declared scalar types (and, indeed, all scalar types) are inherently
ordered. In other words, there is a lowest value, a highest value,
and some number of distinct values in-between. I n your data type
Months, January is the lowest value, December is the highest
value, and February through November are the values in-
between. Pascal will let you use these values just as you would
integers, with one major exception: you can't do arithmetic with
them. It's illegal (and meaningless) to write a statement like
Shipelass:= Reliant + Enterprise;. It is, however, perfectly legal to
use DSTs in the following manner:
for ShipClass := Constitution to Baker do ...
if ShipClass >= Excelsior then ...
case ShipClass of
NoShip
Constitution,Enterpr1se
Reliant
Loknar,Chandley
Larson,Baker
Excelsior
end;
In other words, DSTs can be treated almost like numeric values.
As a matter of fact, each element in a DST has an implied numeric
or ordinal value. The very first element has an ordinal value of 0;
each following element then has a value one greater than the
element before it. For example, here are all the elements of your
DST ShipType, along with their ordinal values:
12-4
Element Ordinal Value

NoShip a
Constitution 1
Enterprise 2
Reliant 3
Loknar 4
Larson 5
Chandley 6
Excelsior 7
Baker 8
Pascal provides some functions to help you use this ordered
relationship. Given a variable or constant value of some DST, you
can find (1) the elementthat comes before it (its predecessor), (2)
the element that comes after it (its successor), and (3) its numeric
(ordinal) value. The first function is called Pred; the second, Succ;
and the third, Ord. For example, let's say that you've set
ShipClass: =Reliant. These functions will now yield the following
results:
Pred(ShipClass) Enterprise
Succ(ShipClass) Loknar
Pred(Pred(ShipClass)) Constitution
Succ(Succ(ShipClass) ) Larson
Pred(Succ(ShipClass) ) Reliant
Succ(Pred(ShipClass) ) Reliant
Ord(Pred(ShipClass)) 2
Ord(ShipClass) 3
Ord(Succ(ShipClass) ) 4
Note that for all scalar types except Integer, Ord returns a value
greater than or equal to 0, since all scalar types except integer
start with an ordinal value of O. For values of type Integer, Ord
returns the actual integer value, which can fall anywhere in the
range -32768 through 32767. Of course, using the Ord function
on any Integer (or Byte) expression is a waste oftime anyway, so
the issue is not a critical one.
12-5
Here's another example. You've seen that you can use DSTs in
for statements, but, if you wanted to, you could build your own
loops:
ShipClass := NoShip;
repeat
ShipClass:= Succ(ShipClass);
untU ShipClass = Baker;

This loop would execute once for each element from Enterprise
to Baker. If you wanted to go in reverse order, you could write
ShipClass := Baker;
while ShipClass > NoShip do begin
ShipClass:= Pred(ShipClass)
end;
which would start at Baker and go downwards to Enterprise. Of

course, you could achieve the same effect with
for ShipClass : = Baker downto Enterprise do begin
end;
There is a subtle point to notice in your definition of the DST

ShipType. Even though you are dealing with eight types of ships,
you've defined nine elements; the extra one is your first element,
NoShip. When you define a DST, you should consider putting in
the local equivalent of 0; that is, some element that essentially
means "none of the above". There are two good reasons for this.
First, you will find yourself in situations where you really are
dealing with none of your regular values. Forexample, you might
have a list of ships with which you are attacking:
12-6
var
Ship : array [1 ..10] ofShipType;
(If you're not familiar with what an array is, don't worry-you'll
run into them later. Once you know what they are, come back and
look at this again, and it'll all make a lot more sense.)Suppose that
Ship[5] := Reliant, but that during the course of the game, that
ship gets destroyed. How do you keep track of that information
within the program? You have to come up with some means of
remembering that Ship[5] is gone. If you've included NoShip in
your definition of Ship Type, then you can just set Ship[5] :=
NoShip, and presto! it disappears. This leads to the following
loop, which will ignore all destroyed ships (assume Indx is oftype
Integer):
ifShip[Indx] <> NoShipthen begin
end;
The code in the center of the loop will be executed for each
existing ship; all others will be ignored.
The second subtle reason of a "none of the above" element was
demonstrated earlier, when you created your own loops with
while and repeat statements. I n both cases, you either started or
ended with ShipC/ass equal to NoShip; in other words, NoShip
acted as a "doorstop" of sorts. Without it, you would have had a
much harder time setting up the loops without going out of range,
either with Succ(ShipC/ass), where ShipC/ass = Baker, or Pred
(ShipC/ass), where ShipC/ass = Constitution. The result is either
a range error (if range checking is turned on) or an undefined
value for ShipC/ass (if range checking is turned off). Either case
can cause problems.
There are times when you don't need (or want) a "none of the
above" element. Take, for example, your DST Months. If you
confine your loops to for statements, then you don't need to
define a NoMonth element. On the other hand, it could still be
12-7
useful for two reasons. First, it would allow you to have a true
"undefined" date, rather than just some default date (1/Jan/OO)
that was understood to be undefi ned. Second, it would force the
ordinal values of the elements to match the conventional values
used; Ord(Apri/) would equal 4 instead of 3. The lesson, then, is
think carefully about how you define your DSTs.
TURBO Pascal gives you an additional feature for working with
declared scalar types, one not found in Standard Pascal. You can
convert any scalar type to any other scalar type with the same
numeric (ordinal) value. Forexample, you know that Ord(Re/iant)
give you the Integer value 3; by the same token, ShipType(3)
returns the ShipType value Reliant. Likewise, Boo/ean(NoShip)
returns Fa/se, and ShipType(Fa/se) returns NoShip, since
Ord(NoShip) = Ord(False).
There are two things that you would probably like to do with
DSTs, but which you can't: directly read and write them. For
example, the statement
WriteLn('Ship Class: ' ,8hipClass);
will produce an error when you compile it. The same is true for a
statement such as
ReadLn(ShipClass);
There are, however, some ways around this limitation. If you are
interested in writing out elements of a DST, you will need to
create an array (list) of strings (yes, we'll talk more about both
strings and arrays later). You could do the following:
program example3;
type
ShipType = eNoShip,Constitution,Enterprtse,Reliant,
Loknar,Larson,Chandley,Excelsior,Baker);
var
ShipClass : ShipType;
ShipName : array [ShipType] ofstriDg[12];
12-8
begin
ShipName[NoShip] .= 'No ship'j
ShipName[ Constitution] := 'Constitution' j
ShipName[ Enterprise] .= 'Enterprise' j
{etc.}
ShipName[Baker] .= 'Baker'j
WriteLnC'Ship Class: ' ,shipName[ShipClass])j
end.
Now you can write out the "value" of Shipe/ass by printing the
appropriate string in ShipName. Similar arrays can be set up for
each DST that you want to be able to write out. If you have several
lists like this, you may want to store the strings out in a text file
(and you'll read about those later on, as well) and then read them
in at the start of the program. In that case, your code might look
like this:
var
ShipClass :ShipType;
ShipName:array [ShipType] ofstriDg[12];
InFile :Text;
begin
Ass1gn(InFile,'SHIPNAME.DAT');
Reset(InFile);
for ShipClass := NoShip to Baker do
ReadLnClnFile,shipName[ShipClass]);
Close(InFile);
end.
This assumes that you have a text file SHI PNAME.DAT out on the
disk with the contents:
No ship
Constitution
Enterprise
(and so on. .. )
12-9
Using a separate data file has two advantages. First, it reduces the
size of the source and executable files, often by quite a bit if you
have long arrays to ilitialize. Second, it allows you to change the
string associated with each element without having to recompile
the program.
You have several options for reading in DST values. The first is to
make use of the array you've set up for writing the DST value out
(if, indeed, you have set up such an array). You prompt the user
for a string and search though the array until you've found it:
var
TempStr: str1Dg[ 12];
Found : Boolean;
{ and the rest of the declarations }
begin
Write('Enter Ship Class:'); ReadLnCTempStr);

Found:= False; ShipClass := Baker;
while DOt fOWld and (ShipClass > NoShip) do
if ShipName[ShipClass] = TempStr
then Found:= True
else ShipClass := Pred(ShipClass);
end.
Here, you start at the end of the array and work backward to (but
not including) NoShip. When you find the matching string,
Found gets set to True, and you fall out of the loop with ShipC/ass
set to the correct value. If ShipC/ass gets to NoShip, then you fall
out of the loop with Found = False and must take appropriate
action. You could, for example, put this code with the loop
repeat
until Found;
which would force the user to keep entering the string until one
matched.
12-10
This is a nice, straightforward method, and one that meshes

nicely with your output routine. The only real problem is one of
matching cases (as in "UPPER" and "Iower"). Forexample, if the
user enters 'RELIANT', but ShipName[Reliant] ='Reliant', then
the two aren't going to match, and Found will remain False.'So,
you'll need to put some code in right after" ReadLn(TempStr);" to
convert TempStr's case to match that ofthe strings in ShipName.
In yourexample, you would need to convert TempStr[1] to upper
case and the rest of TempStr to lower case.
If you don't want to go to all that trouble-and especially if you
don't want or need ShipName for output-then there are other
ways of reading in values and converting them to DSTs. You
could, for example, take advantage of the retyping functions in
TURBO Pascal by having the user enter the ordinal value of the
element desired, then converting it to the appropriate element:
var
Indx :Intege~
Sh1pClass :Sh1PJ?ype;
begin
WrlteLn('Sh1p Classes:');
for Sh1pClass:= NoSh1pto Baker do
WrlteLn(Sh1pName[Sh1pClass],': ',Ord(ShipClass));
repeat
Wrlte('Enter value: '); ReadLn(Indx)
untU(Ord(NoSh1p) <= Indx) and(Indx <= Ord(Baker));
Sh1pClass := Sh1PJ?ype(Indx);
end.
This example assumes that you did set up ShipName but decided
to use numeric (rather than string) input. It first writes out the
names of the different ship classes, along with their ordinal
values. It then goes into a repeat loop which won't let the user out
until a correct value is entered. Finally, it converts that value tothe
appropriate ShipType element and assigns it to ShipClass.
12-11
There are other methods, such as prompting for a single

character, then using a case statement to convert from the
character to the DST element, but we'll leave those for now. The
upshot is this: declared scalar types can tremendously aid
program development, documentation, and maintenance, espe-
cially if you think carefully in how they are declared and used.
12.1 SUBRANGES
Besides creating brand-new scalar types, you can also define

subranges of existing (or newly-declared) scalar types. TURBO
Pascal already defines one such subrange: the type Byte, which
is a subrange of Integer. To declare a named subrange, you
simply use the form:
ty:pe
SubRange= FirstVal..LastValj
where FirstVal and LastVal are values of a scalar type such that
FirstVal is less than Lastval. The data type SubRange can now
legally have any value from First Val through LastVal. Subranges
can also be directly declared, that is, a given variable can be
directly declared as having a subrange value, as opposed to
having to first declare the subrange. Here are some examples of
subranges (with some accompanying declarations):
CODBt
XMax = 8j
YMax 10j
Sh1pMax 40j
ty:pe
Daors (Mon,Tues,Wed,Thurs,Fri,8at,8un)j
Sh1pType (NoS h1p,Constitution,Enterpr1se,Reliant,
Loknar,Larson,Chandley,Excelsior,Baker)j
Cruiser Constitution.,.Reliantj
Nibble 0 ..16j
UpperCase '.A! ..'Z'j
Lowercase 'a' ..'z'j
WeekDaors Mon..Fri;
WeekEnd Sat..8un;
XRange 1.,XMaxj
YRange 1..YMaxj
12-12
var
SX XRange;
BY YRange;
ShipCount O..ShipMa.x;
Alphas .. ,
"'-'.
Subranges are often used as index ranges for arrays, as data
types for both arrays and records, and as part of the declaration
of a constant set. In all these functions, they serve to limit the
amount of RAM used and to set boundaries as to what values can
be used.
12.2 DIRECT DECLARATIONS
The example above showed that you can declare subranges

directly; in other words, you can write
var
ShipCount :O..ShipMa.x;
rather than
type
ShipRange =O ..ShipMa.x;
var
ShipCount :ShipRange;
Being able to directly declare subranges like that saves you from
having to come up with names for every subrange that you want
to use. What you may not realize is that you can do the same for
declared scalar types, and, indeed, for any data type (array,
string, record, set, pointer, file). However, there are a few
restrictions on doing this. First, all of the variables of that data
type have to be declared at the same time and in the same place.
You can't write
: (Mon,Tue,Wed,Thu,Fr1,Sat,8un);
: (Mon,Tue,Wed,Thu,Fri,Sat,8un);
12-13
You'll get an error when the compiler hits the second declaration,
since the second set of days uses the same identifiers as the first
set. Instead, you would have to write
var
D8iY'l,DBJT2 : (Mon,Tue,Wed,Thu,Fri,Sat,8un);
Second, you won't be able to pass any such variables as

parameters to a subroutine or procedure, since you won't have a
data type as which to declare the formal parameter. (There is a
limited exception to this-untyped var parameters-but that's a
special case.) Finally (and this is specific to DSTs) you won't be
able to do any retyping, since there isn't a DST identifier to use.
Well, this is the end of the chapter on scalars. If you take a look at
the program diskette supplied with TURBO Pascal Tutor, you will
find a program called SCALARS.PAS. You should study this
source code and make sure you understand how it works. Then
compile it (by using yourTURBO Pascal diskette, of course) and
see how it runs.
12-14
Part 2 Chapter 13 Arrays
·13. ARRAYS
In previous sections, you've learned about the five predefi ned

data types-Integer, Byte, Real, Boolean, and Char-as well as
declared scalar types (DSTs). A variable of one of these types can
hold only one value at a time. For example, if you define:
Val'
Index : Integer;
then Index has only one specific value at any moment. However,
there are situations where you'd like to have a list of values, such
as a list of numbers or characters. That's where arrays come in.
We've talked some about the "Trek" game, so that will be the
examp!e here. Let's suppose that the game area is an 8 by 10 grid.
Each square in the grid is called a sector. Each sector contains (1)
oor more stars, (2) 0 or more enemy ships, and (3) possibly one
starbase. Now, with what you've learned in previous chapters,
how would you represent the grid?
Tough, huh? Now you see why we need arrays. Simply put, an
array is a collection of variables of identical type, each one of
which may be referenced by a unique index value. For example:
var
List : array[ 1 ..10] of Integer;
The array List is a collection of 10 integer variables, namely,
List[1], List[2], List[3], List[4], List[5], List[6], List[7], List[B],
List[9] , and List[10]. You can use any of those 10 variables
anywhere you could use any integer variable. Furthermore, the
index value doesn't have to be a literal value ('1' through '10'); it
can be any expression that resolves to an integer in the range
1.. 10. For exam pie, the statement
for Index := 1 to 10 do
List[Index] := 0;
would set each of those ten variables to O.
13-1
Chapter 13 Arrays Part 2
An array definition takes the form

array[ index range] of data. type;
The index range can be an implicit or explicit subrange of any

scalar data type (integer, char, boolean, DST), subject, of course,
to certain limits. For example, it is extremely unlikely that you
would want to define
var
BigArrBiY : array[Integer] of Char;
since that would allocate space for some 65,000+ characters; the
implied subrange being -32,768 ... 32767 for TURBO Pascal.
Implicit subranges are usually used with DSTs; for example:
'tyJIe
D~ (Sun,Mon,Tues,Wed,Thur,Fri,Sat);
var
Regular array[Mon..Fri] of Integer;
Overtime array[D~] of Integer;
Present array[D~] of Boolean;
The array Regular has an explicit index range (Mon .. Fri), while
the arrays Overtime and Present have an implicit index range
(Sun ... Sat).
The data type of any array can be almost any data type- Integer,
Boolean, Char, Real, DST, arrays, records, sets, pointers, strings,
oreven files. In fact, arrays with multiple index ranges (known as
multi-dimensional arrays) are just arrays of arrays (of arrays ... ).
For example, you might define your game information as follows:
var
Stars,8hips,Base: array[1..8] of array[l..lO] of Integer;
To reference the elements of such an array, you can write
statements like these:
Stars[3][2] := 0;
if Base[X] [Y] =0
then WriteLnC'N0 starbase present')
else Refu.elShip;
Danger:= 5 * Ships[X] [Y];

13-2
In writing this, you are selecting an array, then a particular

element of that array; for example, Ships[X] [Y] refers to the Yth
element of the Xth array. Both notations can get a little tiring,
especially for arrays with three or more index ranges, so Pascal
allows the following shorthand:
var
Stars,8hips,Base : array[ 1 ..8,1 ..10] of Integer;
Stars[3,2] := OJ
if Base[X,Y] = 0
then WriteLn('No starbase present')
else RefuelShipj
Danger:= 5 * Ships[X,Y]j
This makes for Simpler code, but it can hide the fact that you are
working with arrays of arrays. Consider another solution for your
game using a 3-dimensional array:
type
SectItem = (Stars,8hips,Bases);
SectArrBiY = array[l..8,1..10] of Integer;
var
Sector: array[SectItem] ofSectArrBiY;
Temp SectAl'rBiYj
Now, YOUCould refer to the number of stars in Sector 3-2 as
Sector[ Stars, 3,2] , Sector[ Stars] [3,2], or Sector[ Stars] [3] [2]. What
you might not realize is that you could also do the following:
Temp:= Sector[Stars];
This single statement copies all of the values Sector[Stars,1,1]
through Sector[Stars,8, 10] into the locations Temp[1,1] through
Temp [8, 10]. Why? Because both Sector[Stars] and Temp are
defined as being of type SectArray, and TURBO Pascal will allow
direct assignment of identical array types. In fact, any operation
you could do on Temp, you could also do on Sector[Stars] ,
Sector[Ships], and Sector[Bases]. The key to making this work is
to define an array type-such as SectArray-and then using that
in all the appropriate declarations.
13-3
This approach has some real advantages, but other consider-

ations are needed. For example, each and every location in
Sector is an integer, 16 bits' worth of information yielding over
65,000 values. But you don't really need all that space in each
location. You only need one bit (yes or no) for base information,
and only a few bits more each for ships and stars. So, if you're
concerned about space, this might be a better solution to the
problem:
ccmst
ShipMax 15;
StarMax 9;
XMa.x: 8;
YMa.x: 10;
type
XRange 1 ..XMa.x:;
YRange 1 ..YMa.x:;
var
Stars : array[XRange,YRange] of O..8tarMa.x:;
Ships : array[XRange,YRange] of O..8hipMa.x:;
Base : array[XRange,YRange] of Boolean;
This approach has two advantages. First, it prevents you from
getting an "illegal" value for the numbers of stars, ships, or bases
in a given sector. If Ships[3,2] is ever less than 0 or greater than
ShipMax, you'll get a range error, which will tell you that
something has gone wrong somewhere. Second, it will reduce
the amount of memory required for your arrays. The array Sector
requires 480 bytes of memory (RAM), while the arraysStars,
Ships, and Base require a total of just 240 bytes -only half the
amount. The difference may not seem like much, but if you have
lots of arrays-or if your arrays are large-the savings can be
significant.
Another (less preferred) way of indexing is to use the {$R+}
compiler option to flag "out of range" indexing errors. The
drawback to this method is that it uses more space and slows
down compilation.
Yet another solution would be to define an array of records;
however, that will have to wait until you reach the appropriate
section in this tutorial.
13-4
13.1 PACKED ARRAYS

The discussion on storage space brings up another issue.
Standard Pascal defines defines two kinds of arrays:regular
arrays and packed arrays. It also provides two procedures, Pack
and Unpack, to convert between the two types. TURBO Pascal
doesn't distinguish between the two, though it allows (and
ignores) the keyword Packed. Instead, TURBO Pascal auto-
matically attempts to pack all arrays, that is, store them in the
smallest possible space. The lowest space used is one byte per
element; for example, the array Base sets aside one byte for each
location, even though it technically needs only one bit. Why the
extra space? It allows any element of an array to be passed as a
var parameter to a procedure or function. Other Pascal imple-
mentations don't allow elements of packed arrays to be used in
that manner. One more point: since they aren't needed, the
procedures Pack and Unpack aren't defined.
13.2 ARRAY INITIALIZATION

Often you will want to initialize an array, setting all of its elements
equal to a single value. For example, suppose that you wanted to
set all of the elements in Base to Fa/se, so that you could later set
just the ones desired to True. One way of doing this would be:
for X := 1 toXMax do
forY:= 1 toYMax do
Base[X,Y] := False;
Of course, this takes a while to do and uses up a bit of space for
code and variables. On the other hand, you could make use of
two built-in TURBO Pascal procedures. The first is FiIIChar,
which takes the format
FillChar(Dest,Length,Data);
where Pest is the variable (of any sort) to be filled, Length is the
number of bytes to initialize, and Data is the value to which to set
each byte (and can be expressed either as a character or as a byte
value). You know what you want to fill-Base-and you know
13-5
what you want to fill it with-O, which is Drd(False), that is, the
numeric equivalent of False. Now you just need the length in
bytes ... which brings you to the next procedure: SizeD'. SizeD'
can take as its argument any variable or the name of any data
type.lt returns the size of that variable (or of a variable of that
type) in bytes. So, to initialize Base, you could write:
FillChar(Base,8izeOf(Base) ,0);
This statement will set all bits and bytes in Base = to o. The
combination of FiIIChar and SizeD' is a hard one to beat,
especially for array initialization.
13.3 ORDER OF ELEMENTS
The elements of an array are stored in a specific order. The order

is different for CP/M-SO systems than for all other version of
TURBO Pascal. The following description applies to all versions
of TURBO Pascal except CP/M-SO:
If the array is one-dimensional-that is, if it has only one index-
then the elements are stored in ascending order. For example,
the array List (defined as array[O.. 9] of Integer) stores its
elements in the order List[O], List[1], List[2], and so on ... basically
what you would expect. But what about multi-dimensional
arrays? The array Stars is defined as
var
Stars : array[XRange,YRange] of O..StarMa.x;
where SRange = 1... S. So, the question is, are the elements in
Stars stored as Stars[1, 1], Stars [2, 1], Stars[3, 1], etc., or are they
stored as Stars[1,1], Stars[1,2], Stars[1,3], and so on? Pascal
itself gives you the answer to that question. Remember that your
definition above is just shorthand for
array[XRange] of array[YRange] of O..Starma.x;
13-6
In other words, Stars' first index doesn't select an element, it

selects an array[YRange] of O.. Starmax. The second index
selects an element within that array, and those elements are
stored sequentially, just as in List. So Stars[1, 1] says to pick the
first element of the first array; Stars[1,2], the second element of
the first array; and so on. So the elements in Stars are stored in the
order
Stars[l,l]
Stars[1,2]
Stars[1,3]
Stars[1,4]
Stars[1,5]
Stars[1,6]
Stars[1,7]
Stars[1,8]
Stars[1,9]
Stars[1,10]
Stars[2,1]
Stars[2,2]
Stars[8,9]
Stars[8,10]
All you need to do is remember that the index furthest to the
right-the last index-changes the fastest. If you have the array
var
BigOne : array[0 ..3,0 ..4,0 ..6,0 ..2] of Byte;
then you can quickly work out that the elements are stored as
BigOne[O,O,O,O]
BigOne[ 0,0,0,1]
BigOne[O,O,O,2]
BigOne[ 0,0,1,0]
BigOne[3,4,5,1 ]
BigOne[ 3,4,5,2]
13-7
The above description generally applies to CP/M-80 systems,

with the following important difference:
The elements of the array are stored in descending rather than
ascending order. When you address the array, the first element is
located just as in other versions; however, additional elements
will be found in descending memory locations rather than in
ascending memory locations
13-8
Part 2 Chapter 14 Strings
14. STRINGS
-String-
When Niklaus Wirth designed Pascal, he did so in a punched-
card/mag tape/mainframe environment, where fixed-length data
were the rule. At least, that's probably the reason he was satisfied
to store a character string as a array[1 .. n] of Char. At any rate,
Standard Pascal does not (currently) have a predefined data type
for strings. String constants such as
ccmst
FlleName = 'B:STABB.DAT';
LifeName = 'WHATEVER YOU WANT';
are considered to be array[1 .. n] of Char, where n equals the

number of characters in the string.
The early mainframe (that is, very large) computers were batch-
oriented systems. "Jobs" were "submitted" using large decks of
punched cards or reels of magnetic tapes, and the results came
14-1
Chapter 14 Strings Part 2
out on a high-speed line printer. But the arrival of interactive

operating systems ("timesharing") and the CRT terminal started
a new approach to computer use. When the minicomputer
showed up, so did the first code designed to interactively
manipulate text: word processing programs. The explosive
growth of the microcomputer market over the last 10 years has
been matched by an equal growth in word processors and the
number of people using them. For all its reputation as a number-
crunching machine, the computer is used most often to move
words, not values.
The basic concept behind text manipulation is that of a string. A
string is simply a sequential list of characters of some length. For
our purposes, the characters belong to the ASCII character set
(the back of the TURBO Pascal Reference Manual gives you the
list). The string can contain letters, digits, and pun9tuation.lt can
even have non-printing (control) or special characters. You can
pick out parts of the string, add to it, take away from it, combine it
with other strings, print it out, read it in--i'n short, you can
manipulate it.
TURBO Pascal allows you to declare a variable to be of type
string, followed by a length specification. For example, you could
define the following: "
var
MyName striDg[ 80];
Token striDg[ 15];
BlgString striDg[ 255 ];
Note that you must specify a length for each variable. This
defines the maximum number of characters that each string can
hold. The variable MyName could hold up to 80 characters.
Token could only hold up to 15 characters, so that the statement
Token := 'this is too long a string for token';
would only store the first 15 characters ('this is too Ion') into
Token. The last variable, BigString, represents the maximum
length possible for a string-255 characters.
14-2
The data type string[n] can be thought of as an array[O .. n] of

Char. For example, you can reference individual characters in
Token using the notation Token[1], Token[2], and so on. The first
location, Token[O], contains the current length of Token. If you
execute the statement
Token:= 'this string';
then Token[O] contains the value 11, since there are 11 characters
in'this string'. However, you could not do something like this:
program Example;
var
Token : striDg[15];
Len : Integer;
begin
Token: ='this string';
Len := Token[O];
WrlteLn('The length of token is ',Len)
end.
Why not? Because Token[O] is of type Char, and you can't assign
a character to an integer. You could, however, substitute the
statement
Len:= Ord(Token[O]);
which would return the ordinal (numeric) value of Token[O],

which happens to be 11.
Using the array notation, you can play with any individual
character of a string. As mentioned above, each element of a
string is a variable of type Char, and you can treat it as such. For
example, you might want a procedure to convert all letters in a
string to uppercase ('A' .. 'Z'):
14-3
procedure LowToUp(var Str: striDg[255]);

{
purpose converts characters in Str to upper case
last update 27 Oct 85
var
Indx,Len : Integer;
begin
Len:= Length(Str); { more on this later}
for Indx := 1 to Len do
Str[Indx] := UpCase(Str[Indx]) {bunt-in TURBO func}
end.; { of proc LowToUp }
A caveat (warning) is in order. You should avoid trying to mess

with any elements beyond the current length of the string.
TURBO Pascal won't give you any sort of error, but you need to
be aware that you've just changed a portion of the string that
won't print out unless you change the length as well. (Note that if
you use the {$R+} compiler directive, and error message will
display.)
You've already seen that you can assign string constants to string
variables (Token := 'this string'). You can also read and write
strings through textfiles (which will be discussed in a later
section). Most notably, you can read and write strings through
the predefined textfiles, Input and Output. For example, this
program allows you to type in a line (up to 80 characters!) and
then writes it back out to the screen. It continues to do this until
you type, "1 quit!":
program Echo;
var
Line : striDg[ 80];
begin
WrlteLn(,Entertng echo mode-type"I quit!" to exlt');
repeat
ReadLn(Line );
WrlteLn(L1ne)
until Line = 'I quitl'
en4 { of program Echo}
14-4
Now, let's make the following modifications to your program

(don't worry about the file stuff; yes, you'll learn about that later,
too):
program MicroWord;
var
Line : striDg[ 80];
OutFile : Text;
begin
Write('Enter file name: '); ReadLn(Line);
.Assign( OutFlle,Line); Rewrite( OutFile);
WriteLn('Entering inSert mode-type "I quit!" to exit');
repeat
ReadLn(Line);
if Line < > 'I quitJ'
then WriteLn( OutFile,Line)
until Line = 'I quit!';
Flush(OutFile); Close(OutFile)
end. { of program MicroWord }
Voila! You've just written a word processor. However complex or

sophisticated they may seem, all word processing programs
eventually boil down to the program above. Start with this
program, add modifications, and eventually you'll have your own
text editor.
14.1 STRING COMPARISONS
Just like numbers, strings can be compared to each other. In the

program MicroWord, you checked to see whether the string
variable Line was equal to the string constant '1 quit!'. The
comparison is simple. First, the lengths of the two strings are
compared. If they're different, then the strings are not equal. If
they're the same, then the characters in the two strings are
compared, starting with the first one and continuing until (1) two
characters are different or (2) all characters have been compared.
In case (1), the strings are not equal; in case (2), they are. You
could even write a function to show this comparison:
14-5
fanctionStrEqual(Strl,8tr2: BigStr): Boolean;

{
purpose show how strings are compared for equality
Note well:this function is *not* necessary,
since "Strl = Str2" will perform the same
comparison.
var
Len,Indx : Integer;
Flag : Boolean;
begin
StrEqual := False;
Len:= Length(Strl);
if Len = Length(Str2) then begin
Indx:= 1;
Flag := True;
while FlagAND (Indx <= Len) do
ifStrl[Indx] = Str2[Indx]
then Indx := Indx + 1
else Flag := False;
StrEqual := Flag
end
end; { of func StrEqual }
Once again, please understand that you do not need this

function. This just shows how the Boolean expression Str1 = Str2
comes up with a value of True or False.
Similar comparisons occur when you want to see if one string is
"greater than" or "less than" another. For example, let's suppose
you're sorting a list of names into alphabetical order. At some
point, you'll compare two strings to find which comes before (is
less than) the other. The statement
ifStrl > Str2 then ...
will take some action if and only if Str2 comes before Str1. The
comparison algorithm can be described as follows:
1. Point to the first character of each string
2. Compare the two characters
3. If they are not the same, go to 8.
14-6
4. Get the next two characters

5. If they're both there, go to 2.
6. If only one string has characters left, then it is greater than the
other
7. Otherwise, neither string has characters left, so neither is
greater than the other (they're identical!). Stop
8. One character has a greater ASCII value than the other. The
string that character came from is greater than the other
string. Stop
To summarize, you have three cases. First, both strings are the
same length and have the same contents. In that case, they're
equal. Second, both strings have the same contents up to the end
of one string; the other string has additional characters beyond
that point. In that case, the shorter string is less than the longer
one. In the last case, the strings cease to match at some point (it
may even be the first character). In that case, the string whose
unmatched character has the lower numeric (ordinal) value is
less than the other string.
14.2 STRING FUNCTIONS AND PROCEDURES

Another useful aspect the TURBO Pascal string definition is that
it includes more that just the data type string. It also defines a set
of functions and procedures which work on strings. The table at
the end of this chapter lists all of them, but we'll go over each one
at a time (and not necessarily in the order given).
The most commonly-used function is probably Length(St),
which was used in a few of the examples above~ It's really just
another way of writing Ord(St[O]); that is, it returns the current
length of St. This is not to be confused with the maximum
possible length of St. For example, the program
14-7
program LengthTest;
type
SmallStr = strmg[16];
var
Test : SmallStr;
procedure ShowLength(St: SmallStr);
begin
WriteLn('length of<' ,Bt,'> is' ,Length(St))
end; { of proc ShowLength }
begin
Test: = 'hello, there';
ShowLength(Test);
Test := 'hi';
ShowLength(Test);
Test: = ";{ null string }
ShowLength(Test)
end. { of program LengthTest }
will produce the output

length of <hello, there> is 12
length of <hi> is 2
length of < > is 0
The next most commonly-used function is probably Concat. You
can use it to patch several strings together. It's handy for inserting
string variables in the middleoffixed messages. Forexample, this
program
program Concatenation;
var
Name,Message: striDg[30];
begin
Write('Please enter your name:');
ReadLn(Name);
Message:= Concat{'Hello, ',Name,' how are you?');
WriteLn(Message)
end. { of program Concatenation}
14-8
produces this output:

Please enter your name: Deirdre
Hello, Deirdre, how are you?
Besides the explicit function Concat, TURBO Pascal also lets you
piece together strings using the plus sign (+). For example, you
could change the third statement in your program above to read
Message:= 'Hello, ' + Name + " how are you?';
This has exactly the same effect as the Concat function, and you
may find it easier to use.
There are two things you have to be aware of when concatenating
strings. First, as mentioned above, if the resulting string is longer
than the variable can hold, the extra characters will be thrown
away. More importantly, if the resulting string is longer than 255
characters, you'll get a run-time error, and your program will
come to a screeching halt (unless you have used the {} compiler
option, in which case you would have gotten an error during
compilation).
Right up there with Concat is the function Copy, which allows
you to pull out part of a string (called a substring). It takes as
parameters the string itself and the substring's location and
length. The code
Name := 'Deirdre Ann Webster';
Middle := Copy(Name,9,3);
WriteLn( 'Your middle name is ',Middle)j
would produce the output "Your middle name is Ann".

Of course, the example above depended upon knowing right
where 'Ann' started in the string. Suppose you knew what string
you were looking for, but didn't know where it started? You could
use the function Pas to find it. Given a substring or pattern, and
the string in which to search for it, Pas returns the location of the
start of the substring. If it can't find the substring, Pas returns O.
Your code could now look like this:
14-9

Loe:= Pos('Ann' ,Name);
Middle:= CopyCName,Loe,3);
WriteLnC'Your middle name is' ,Middle);
The Delete procedure lets you remove a section of a string. Like

Copy, it requires the string, the starting position, and the number
of characters to delete. For example, you could change the code
above to read
Loe := pose'Ann' ,Name);
Delete(Name,Loe,4); {need to cut out an extra blank}
WriteLn('Your name is ' ,Name);
which would then write out "Your name is Deirdre Webster",

having deleted the string 'Ann '.
Combined with Pas and Copy, Delete can be used to parse a
string, that is, to pull off a chunk at a time. Let's suppose that you
want to write a procedure that will pull the first word off a string,
where a "word" is defined as any substring starting with a non-
space character and followed by a space. Our procedure might
look like this:
14-10
procedure Parse(var Line,Word: BigStr);

{
purpose removes first word in <line> and returns it in <word>
}
cODBt
Space =' '.,
var
Indx,Len : Integer;
begin
while Pos(Space,Line) = 1 do{ remove leading blanks }
Delete(Line,1 ,1 );
Len:= Pos(Space,Line); { look for blank}
if Len = 0 then begin { no blanks left }
Word := Line; { get word}
Line:=" { zero out line}
end
else begin { get word and delete from line}
Word := Copy(Line,1 ,Len-1); { get all but blank }
Delete(Line,1,Len) { delete word plus blank}
end
end; { of proc Parse}
The next procedure, Insert, is the reverse of the Copy/Delete

operation: it takes one string and stuffs it somewhere inside
another. You just specify the string you want to insert, the string
into which it's to be inserted, and the location of the the insertion.
Forexample, if you wanted to put Deirdre's middle name back in,
you could use the following code:
Name := 'Deirdre Webster';
Middle := 'Ann';
Insert(Middle,Name,Pos('Webster' ,Name));
Insert(' , ,Name,Pos('Webster' ,Name));
WriteLn( 'Your full name is ',Name);
Note that you use the Pos function to figure out where to insert
'Ann' and also where to insert a blank (to separate 'Ann' and
'Webster') .
Token expansion is one good use for Insert. Suppose you were
writing a program to take a form letter and put in the appropriate
14-11
names, dates, and so on. Within the form letter, these fileds might
be represented by tokens; for example, the salutation might look
like this:
Dear <title> <last name>:

The program could scan through the form letter, list all the tokens
(which all have the form < ... »,
then get the information to
replace them (interactively, from a file, etc.). The following
procedure, then, might be of use:
procedure Replace(var Line: B1gStr; Token,Sub: TokStr);
{
purpose look for Token in Line and replace with Sub
}
var
Indx,Len : Integer;
begin
repeat
Indx := Pos(Token,Line);
if Indx > 0 then begin
Delete(Line,Indx,Len.gth(Token) );
Insert(Sub,Line,Indx )
end
untll Indx = 0
end; { of proc Replace}
The statements
Line := 'And so, <title> <last>, the entire <last> family'j
Replace(Line,'<title>' ,'Dr.');
Replace(Line,'<last>' ,'Lewis');
WriteLn(Line);
would produce
And so, Dr. Lewis, the entire Lewis fa.m1ly
This should give you a clue on hawaII that "personalized" junk
mail that you receive is generated.
14-12
14.3 NUMERIC CONVERSIONS

TURBO Pascal provides two procedures for converting numbers
to strings and vice versa. These two procedures work in a fashion
similar to Read and Write, which you've had some exposure to
and which will be discussed more in a later section. The first
procedure, Str, will convert a number into a string, formatting it
much as it would for text output. The number can be either
Integer or Real, and you can specify the width format just as you
can for output. Here are some calls to Str, along with the resulting
strings (for our purposes, X = 4.281953E3 and I = 14916; S is
defined as string [12] ):
call to Str contents of S
Stl'( X:12:3,8);" 4281.963"
Stl'( X:12:0,S); " 4282"
Str( X:10:7,8); "4281.9630000"
Str( -X: 12:6,8); " -4281.96300"
Stl'( X: 6:4,8); "4281.9630"
Str( 1:12,8); 14916"
Stl'( I: 6,8); "14916"
Stl'(-1: 7,8); "-14916"
Btl'( I: 3,8); "14916"
Notice that Str does, indeed, behave like numeric output. The
string length is set equal to the field width; if the width is too small
(such as X:1 0:7, X:5:4, or 1:3), it is increased to fit the number. If the
field is wider than is necessary, then the number is right-justified,
that is, blanks are put in front of the number to fill out the
remaining space. In the case of real numbers, rounding off is
done when needed.
The second procedure, Val, converts from a string to a number
(again, either Real or Integer). The string itself must contain
exactly a number and nothing else; no characters other than
digits, except for '+','-','.', and 'E' in the appropriate places. And,
of course, the number in the string must be of the same type as
the variable to which Val is converting it. Since there are so many
chances for error, Val has a third parameter-a result value-
14-13
which tells you whether or not there were any problems. If the
result is 0, then there were no problem during the conversion. If
the result is greater than 0, then it indicates the character
(S[Result]) at which it ran into problems. Here are a few
examples:
!fS holds
the string: then VALCS) = result code
"14916" 14916 0
" 32" <undef> 1 {space}
"4281.963" 4281.963 0
"-332.3" -332.3 0
"-332.3 " <undef> 7 {space}
"4,281" <undef> 2 {comma}
Before doing any heavy-duty numeric conversions, you would be

well advised to read the Formatted Output section of Chapter 18,
to get a good understanding of the concept of field widths.
A word of caution for those of you using 8-bit (CP/M) systems: do
not use Str or Val within a function that is itself called within a
Write or WriteLn statement. Strange and undesirable things will
happen as a result. Instead, call the function beforehand,
assigning its value to some variable, then use that variable in the
Write/WriteLn statement. Those of you with 16-bit machines
(CP/M-86, MS-DOS) needn't worry about any of this.
14.4 STRINGS AS PARAMETERS
You've probably noticed in these examples that whenever we

pass a string to a procedure or function, we've defined that
parameter as something like BigStr or TokStr, rather than
string[255] or the like. For example, we used
procedure ParseC var Line,Word: BigStr);
instead of
procedure ParseC var Line,Word: striDg[266 ]);
In TURBO Pascal, you cannot directly declare a parameter as

being a string of some given length. Instead, you must declare a
14-14
data type that is equivalent to a string of some length, then use

that data type in the parameter declaration, like this:
program ParseText;
type
B1gStr = string[255J;
procedure Parse(var Line,Word: B1gStr);

You have to take one other factor into account. When you call a
procedure or function that has string parameters, and those
parameters are declared as var, that is, they can be changed by
the procedure, then the string variables you pass must be
declared to have the same length as the parameters. For example,
suppose you added a procedure like this to our example above:
procedure DoParsing;
var
TLine,TWord : string[80 J;
begin
Write('Enter line: '); ReadLn(TLine);
WriteLn('Parsed line: ');
while Length(TLine) > 0 do begin
Parse(TLine,TWord);
WriteLn('<' ,TWord,'>')
end
end; { of proc DoParSing }
If you tried to compile the program with this addition, you would
get a type mismatch error when it got to the line Parse(TLine,
Tword);. Why? Because TLine and TWord are declared to be of
length 80, while Parse is expecting two strings of length 255. This,
of course, can cause real problems if you're trying to write some
general-purpose routines (such as LowToUp) to handle all
different strings. The reason for the error is to prevent you from
returning too long a string or indexing into "random" memory
(that is, beyond the end of the string). You can, however, disable
this error checking by putting a compiler option at the start of
14-15
your program (more about these later in the book). All you have
to do is to place the following comment somewhere before the
call to Parse; you could, in fact, put it at the top of your file, like
this:
{$V-}
program ParseText;
TURBO Pascal will then no longer check to see if the string

lengths match. Like most "disable" options, you should use this
with caution; if you aren't careful with passing different length
strings (that is, strings with different defined maximum lengths)
to the same procedure, you could get some bizarre errors. One
method would be to disable the checking for each specific call,
such as
{$V -} {turn off string checking }
Parse(TL1ne,Tword);
{$V +} {turn on string checking}
That way, you will be turning it off for only those places where
you actually do not need it.
14.5 STRINGS, CHARACTERS, AND ARRAYS
TURBO Pascal makes it easy to mix references to strings,

characters, and arrays. You can use strings and characters
interchangeably; in other words, you can use a string almost
anywhere you use a character, and vice versa. There are,
however, a few exceptions. First, you can not use a character as a
parameter to a procedure or function where a var parameter of
some string type is expected. For example, if you have a variable
Ch of type Char, you couldn't write
UpToLow(Ch);
14-16
since Up ToLow wants to pass a string back. The {$V-} option

won't even work (and it's probably just as well). Second, if you
assign a string to a character, the string must have a length of
exactly 1; a longer or smaller (null) string will result in a run-time
error.
In a similar fashion, you can use variables that are declared as
array of Char as strings, again with a few restrictions. You can't
assign string variables to arrays, nor use arrays as var string
parameters. You can assign string constants to arrays if the
constant is exactly the same length as the array.
Here is a program which demonstrates many of the ways in which
you can mix strings, characters, and arrays of characters.
Try to predict ahead of time what your output will look like; then
key it in and run it, and see how close you were.
program Test;
type
B1g8tr = str1ng[ 266];
var
Stl,St2 striDg[ 40];
Chl,Ch2 ChaI';
Value,Code Integer;
.A:rl;U-2 array [0 ..9] ofChaI';
procedure PutLen(St: BigStr);
begin
Wr1teLn(St,': ',Length(St))
end; { of proc PutLen }
14-17
{ main body of program Test}

begin
ChI := 'A'; Ch2 := 'Z';
Stl : = 'The alphabet goes from' + ChI + ' to ' + Ch2;
Arl := '0123456789';
PutLen(Arl );
St2:= 'j';
Chl:= St2;
Ch2:= 'I';
PutLen(Stl );
PutLen( St2);
WriteLn('Chl in St2: ',Pos(Chl,St2));
WriteLn('Ch2 inArl: ',Pos(Ch2,Arl));
Vale Ch2,Value,Code);
WriteLn(Ch2,' ',Value,' ',Code);
Stl :=Ch2;
Ar2:= '0000000012';
Val(Ar2,Value,Code);
WriteLn(Ar2,' ',Value,' ',Code);
St2 :=Arl;
PutLen( St2)
end { of program Test}
Now, here is the table I promised you earlier in the chapter. This
sort of brings all of the string procedures and functions together
into one location, so you can easily compare what they do.
STRING PROCEDURES AND FUNCTIONS
Concat(S1,S2, ... ,Sn) Function which returns string composed of S1
through Sn concatenated together; the plus sign
(+) can also be used
Copy(St,lndex,Size) Function which returns string composed of
St[lndex] .. St[lndex+Size-1 ]
Delete(St,lndex,Size) Procedure which deletes Size characters St
starting at St[lndex]
Insert(St1,St2,lndex) Procedure which inserts St1 into St2 starting
at St2[lndex]
Length(St) Function which returns current length of St
Pos(Pat,St) Function which returns position (index) of Pat
within St
Str(Val,St) Procedure which converts Val (Integer or Real)
into a string and stores it in St
Val(St,Val,lndex) Procedure which converts St into Val (Integer or
Real) and sets Index to the position of any error
occurring (0 = none)
14-18
Part 2 Chapter 15 Records
15. RECORDS
BY8
-Records and Files-

You've learned about arrays, which allow you to create collections
of objects of the same type, indexed by some sort of value. But
what if you want a collection of objects of different types? You
could declare an array for each data type, but that could get
tedious and complicated. Pascal offers another solution:records.
The record data structure is massively useful, if handled correctly.
It allows you to glue together a lot of different data types into a
single structure. The basic definition of a record is:
15-1
Chapter 15 Records Part 2
record
<Identl> : <DataTypel>;
<Ident2> : <DataType2>;
end.;
A record consists of a number of fields (ldent1, Ident2, and so

on). Each identifier can be of any data type or structure, including
integer, boolean, real, char, DST, array, record, set, or pointer. To
refer to a field, you give the variable's name, a period, and then the
field's name, such as:
VarName.ldentl
In the section on arrays, a few attempts were made to define the
information for sectors in the "Trek" game. Let's see if a new
approach, using records, might not work best. Suppose you
define the following:
program. Trek;
ccmst
StarMax 9;
ShipMax 15;
XMax 8;
YMax 10;
type
XRange = l.,XMax;
YRange = 1 ..YMax;
ASector=
record
Ships O..8hipMax;
Stars O..StarMax;
Base Boolean
end;
var
Sector array[XRange,YRange] of AS ector;
Note well that Ships, Stars, and Base are not elements of a
declared scalar type, nor are they separate arrays; they are field
identifiers for the record type ASector. Accordingly, you would
refer to the number of stars in sector 3-2 as Sector[3,2].Stars.
Note also that in setting up this record, you've restricted the
possible values of the fields Stars and Ships to a very small
15-2
subrange of integers, and you've changed Base into a boolean

variable. You thus avoid two problems: (1) having your variables
take on legal but nonsensical values (such as -34 stars), and (2)
using too much memory.
You may remember that in the section on Arrays, you used two
approaches: a 3-dimensional ar'ray of integers that used up 480
bytes, and three separate arrays that used a total of 240 bytes.
This array of records also uses only 240 bytes and has the
additional advantage of holding all of the sector information in
one spot (rather than spread out over three arrays).
This record example is a fairly simple one:just three fields, each
on its own line. Let's work on a more complicated example.
Suppose that you wanted to model each of the stars in a sector as
a solar system, complete with planets. You might make a fairly
simple definition for the planets and systems:
type
PlanetClass (Terroid,Jovoid,ABteroid);
StarClass (O,B,A,F,G,K,M,R,N);
StarSize (SubDwarf,Dwarf,Normal,Giant, SuperGiant);
Planet record
D1st,Angle,Radius,Mass,Gravtty
Real;
Class : PlanetClass
end;
System =
record
SX,sy : Byte; {location w/in sector}
Class : StarClass;
Size : StanSize;
Mass,Radius,Temp,Lummosity
: Real
end;
Not a whole lot of information, but enough to get read-outs on the

"bridge status display." Notice that you can define several fields
ofthe same type at the same time (SX and SY; Dist, etc.). Had you
wanted to, you could have given each one a separate line; it isn't
necessary, but it can help for documentation purposes.
15-3
15.1 THE "WITH" STATEMENT
Let's suppose you have the following definitions:

coDSt
XMax 8;
YMax 10;
ShipMax 16;
StarMax g.,
type
XRange 1.XMax;
YRange l ..YMax;
ASector=
record
Starbase Boolean;
SbX,8bY Byte;
Ships O..ShipMax;
Stars O..StarMax
end;
var
Sector : array[XRange,YRange] of ASector;
You've defined an array of records which you can now use in
your program. If you wished to clear the entire sector of starbases
and enemy ships, you might do this:
for X := 1 to XMax do
for Y := 1 to YMax do begin
Sector[X,Y] .Starbase := Falsej
Sector[X,Y] .SbX := 0;
Sector[X,Y] .SbY := 0;
Sector[X,Y].Ships:= 0;
Sector[X,Y] .Stars := 0
end;
As you can see, this could get very tedious, especially for more
complicated records and/or actions. Luckily, Pascal gives you a
shorthand means of referring to all those fields: the with
statement. This takes the form
with <record id> do <statement>;
15-4
Within the statement, you can refer to all of the fields of the record
id without having to specify it each time. Instead, the base
address of record id is calculated when the with statement is
found, and that address is used to offset all of the fields. This
makes it very handy for use in loops. For example, you could
rewrite your initialization code as:
forY:= 1 toYMax do
with Sector[X,Y] do begin
Starbase := False;
SbX:=O;
SbY:=Oj
Ships:= OJ
Stars:= 0
end;
Note that the with statement is inside of the two for statements.
This is critical, for it ensures that Sector[X, Y] refers the desired
element of the array. If you had written
with Sector[X,Y] do
for Y := 1 to YMax do begin
end;
then the base address would be calculated just once: before you
entered the two loops, and using whatever the values of X and Y
were before the loops started.
You can nestwith statements, and you can also list morethan one
record identifier in a single with statement. For example, you
could write the following:
15-5
type
RecTypel =
record
Fieldl Integer;
Field2 Real
end;
RecType2 =
record
Field3 string [ 20];
Field.4 Boolean
end;
var
Reel RecTypel;
Rec2 RecType2;
begin
with Reel do begin
Fieldl := 32;
Field2 := 17.76;
with Rec2 do begin
Str(Fieldl ,Fiel(3);
Field.4 .= (Field2 > 3.14169)
end;
Fieldl .= Length(Rec2.Field3)
end
end.
or you could modify the main body to read

begin
with Reel,Rec2 do begin
Fieldl := 32;
Field2 := 17.76;
Str( Fieldl ,Fiel(3);
Field.4 .= (Field 2> 3.14169);
F1eldl . = Length(Field3)
end
end.
In the first example, you had to explicitly identify Rec2 when

assigning the first character of Field4 to Field3 of Rec1. In the
second example, you could do a direct assignment.
There are a few pitfalls you much watch for in using the with
statement. First, if you have more than one record defined in a
15-6
given with, or if you have nested statements, make absolutely

sure that the records don't have fields with the same name.
Suppose you changed the definition of RecType2 to read:
RecType2 =
record
Field.! striDg[20]j
Field2 Boolean
end;
Now Rec1 and Rec2 have fields with the same names. If you then
write something like
with Reel ,Rec2 do begin
WriteLn(Field.! )j
WrlteLn(Field2)
end;
which record's fields will be written out? The answer: Rec2, since
it was the last defined in the list. Likewise, if you have nested with
statements, the record last appearing has precedence over any
others. In this example, problems are unlikely to occur sincethe
data types ofthe identically-named fields are different; an attempt
to assign a value will either go to the correct field or will result in a
compiler error. But what if both records had Field1 defined as
type Integer? In that case, you might unknowingly assign a value
to the wrong field. The moral:avoid identically-named fields
unless you're very sure of what you're doing.
A similar problem can occur when you create a field that has the
same name as a variable:
program BadExamplej
type
RecType=
record
Name striDg[26]
Age Integer
end;
15-7
var
Rec : RecType;
Name : striDg[25];
begin
Name := 'J. Michael Browning';
with Rec do begin
Name := 'Bob Trammel';
WriteLnCName)
end;
WriteLnCName)
end.
Can you guess what the output of this program is? If you guessed
Bob Trammel
J. Michael Browning
you're absolutely correct. Again, much like the rules of scope you
learned about in Chapter 11, the last declaration takes prece-
dence. In this case, the with statement acts as a declaration.
Again: be careful with identical identifiers.
15.2 VARIANT RECORDS
Occasionally, you may define a record with redundant or

mutually exclusive fields. For example, consider the following
record definition:
15-8
type
Government = (Federation,Kl1n.gon,Romulan,Trader);
AShip =
record
SX XRange;
BY YRange;
QX,QY Byte;
Energy 0 .. 1023;
Source Government;
Phaser,D1Bruptor,Beam1 ,Beam2
Byte;
Photon 0 .. 15;
Torpedo 0 ..7;
Plasma 0 ..9;
Cloaked boolean
end.;
Let's suppose that Federation ships have Phaser and Photon

weapons; Klingons have Disruptors and Torpedos; Romulans
have Beam and Plasma weapons, along with Cloaking devices;
and Traders have no weapons at all. That means ·that each and
every ship record has four unused fields and, therefore, wasted
space. If you only have a few records, the space won't matter
much; if you have a lot, it can become significant. Yet, the records
are so similar that it seems silly to define three different records.
Furthermore,if you want to put all the ships in a single array orfile,
you have to stay with one record. What do you do?
Well, Pascal allows you to create a record with conditional fields,
called a variant record. To create a variant record, you define a
tag field, which is just a field of some scalar type (often a declared
scalar type). This field goes after all the fixed (non-varying) fields
and is used as the argument in a case statement. For each
possible value of that field, you define the desired exclusive fields.
For example, you might rewrite your definition of AShip as
follows:
15-9
type
Government = (Federation,Kl1ngon,Romulan,Trader);
ABh1p =
record
SX XRange;
BY YRange;
QX,QY Byte;
Energy 0 ..1023;
case Source Government of
Federation (Phaser Byte;
Photon 0 .. 15);
Kl1ngon (D1sruptor Byte;
Torpedo 0 ..7);
Romulan : (Beam1,Beam2 Byte;
Plasma 0 ..9;
Cloaked Boolean);
Trader : ()
end;
For each value of Source, you define the fields that exist for that
case, enclosed within parentheses. Note that even if there are no
fields at all, you still need to put in the parentheses.
You may wonder what good this all is. First, it cuts down on the
size of the record. Why? Because the fields for each variant
occupy the same space. In other words, the fields Phaser,
Disruptor, and Beam1 all reside at the same location in memory,
rather than each having their own separate slot. The result is less
memory required for each record of that type. The original
definition required 15 bytes per record; this variant record takes
up only 11 bytes. That may not seem like a whole lot, but consider
this example:
15-10
type
Government Federation,Kl1ng,Romulan, Trader);

Sides (Forward,FPort,FStarboard,APort,
ABtarboard,Att);
Weapons array[ Sides] of Byte;
ABhip =
record
SX XRange;
SY YRange;
QX,QY Byte;
Energy 0 .. 1023;
Federation (Phaser Weapons;
Photon array[Sides] of 0 .. 15);
K1.1ngon : (Disruptor Weapons;
Torpedo array[Sides] of 0 ..7);
Romulan : (Beaml,Beam2 Weapons;
Plasma array[Sides] of 0 ..9;
Cloaked Boolean);
Trader : ()
end;
You've replace a single byte value with an array of six bytes in

several places. A variable of this type now takes up 26 bytes.
However, if you declared this as a non-variant record, it would
require 50 bytes, or nearly twice as much room. If you had an
array or file of these records, the difference could be very
important, indeed.
There is a second, more esoteric aspect of variant records. Since
certain fields now occupy the same space, you can set the value
of one field and have it transfer over to another. As mentioned
above, the fields Phaser, disruptor, and Beam1 all occupy the
same space. Because of that, we could do the following:
var
ThisShip : ABhip;
begin
with ThisShip 40 begin
Phaser[ 1] : = 72;
WriteLn(Disruptor[ 1])
end
end..
15-11
which would write out "72". What good is this capability? Well, it
allows for certain advanced programming tricks. For example,
assuming an a-bit system, you could define something like this:
type
B1g8tr = striDg[ 255];
StringPo1nter=
record
case Flag: Boolean of
True (Addr : Integer);
False (Ptr: A B1g8tr)
end;
var
VStr StringPo1nter;
begin
VStrAddr := $BOOO;
VStr.Ptr A := 'This string is beingwrltten at $BOOO'
end
Now, you can write a string at any given location in memory. Of
course, since you don't always know what is sitting at a given
memory location, this can be dangerous. That's why it's for
advanced programming efforts.
Here's one more variation on variant records. Sometimes you
want the variations, but you don't really care about the tag field.
The StringPointer above is a good example. You're interested in
Addr and Ptr, but you don't need or want Flag. In those cases,
Pascal allows you to dispense with the tag field altogether. For
example, you could rewrite StringPointer as:
type
B1g8tr = striDg[ 255];
StringPo1nter =
record
case Boolean of
True : (Addr : Integer);
False : (Ptr : A BigStr)
end;
Note that you no longer have the tag field, Flag; you just have the
tag field type, Boolean. This is known as a free union (as opposed
to a discriminated union, one with a tag field).
So much for records. ~ '40W, let's move on to sets ...
15-12
Part 2 Chapter 16 Sets
16. SETS
There are times when you want to test a scalar variable (Integer,
Byte, Char, Boolean, DST) to see if its current value belongs to a
set or collection of values. For example, suppose you want to
write a subroutine that will write out a prompt for the user and
then accept and return only a character belonging to an
allowable set of characters. Using what you know now, how
would you do that?
If you're really clever, you might come up with something like
this:
type
CharSet array [ Char] of Boolean;
var
OKSet CharSet;
You could then set each location in OKSet to the appropriate

value (True or False) and use that to check the characters read in.
Unfortunately, this approach tends to eat up lots of memory (128
bytes per CharSet) and program space. Fortunately, it's not
necessary. Pascal handles sets foryou by letting you define sets.
A set can be defined for any scalar data type, though there is one
restriction: the numeric (ordinal) values of the data type (or
subrange of a data type) used must range between 0 and 255. For
example, you could not define a set of Integer, but you could
define a set of Byte, or even a set of 21 ..47. This restriction also
means that you can never have more than 256 items in a set.
A set constant is enclosed by brackets ("[" and "]"), with the
elements of the set defined within. You can list each element
separately; however, that could get tedious for sets with lots of
elements (remember, you can have up to 256!). If a number of the
elements are in contiguous (numerical or ordinal) order, you can
use subrange notation as a type of shorthand. Here are some
exam pies of sets:
16-1
Chapter 16 Sets Part 2
[] { empty set- contains nothing }

[1,3,6,7,9] {set of Byte }
['A' ..'Z'] { set of Char }
[Mon,Tue,Thur] {set ofDaors }
[Jan..Jun.,Aug,Oct..Dec] {set of Months }
[Loknar,Chandley] {set ofShipType}
As mentioned, a set can hold up to 256 different elements, all of

the same data type. It stores the presence of each element as a
single bit; as a result, a set can be at most 32 (256/8) bytes and is
not always that large.
The usefulness of sets is not always obvious. In fact, it can be
hard at times to think of anything to do with them. However,
here's an example that may show one very good use for sets:
program Charrest;
{ $V-} { to avoid any problems passing strings}
type
CharSet = set of Char;
Prompt = str1ng[ 80];
var
Cmd : Char;
procedure GetCha.r( var Ch : Char; Msg : Prompt;

OKSet : CharSet);
begin
repeat
Write(Msg); ReadLn(Ch);
Ch:= (UpCase(Ch); {force to be upper case}
untU Ch in OKSet
end; { of proc GetChar }
{ main body of Charrest
begin
repeat
GetChar(Cmd,'Charrest> S)peak, C)ount, Q)u1t: '
['S' ,'C' ,'Q']);
caseCmdof
'S' : WriteLn('Woofl Woofl');
'C' : WriteLn('l, 2, 3, 4, 6, 6, 7, 8, 9, 10')
end
umU Cmd = 'Q'
end. { of prog Charrest }
16-2
The procedure GetChar is extremely useful for input processing.

It prompts the user with a message, then accepts a single
character response. It converts the character to uppercase, then
checks to see if it's a valid command. If not, it continues to prompt
until a correct selection is entered. To see how and why this
works, you need to understand about set comparisons.
16.1 SET COMPARISONS

The key of GetChar is the set comparison in, which takes the
form:
<element> in <set>
The term element must be an expression of the same type as set's
base type. If set is a set of Char, then element must resolve to a
single character. This expression returns True if and only if
element is currently in set; otherwise, it returns False. In this
example, Ch in OKSet returns True if the value of Ch is in the set
OKSet. Here, Ch must be S, C, or Q in order for this expression to
be True. Note that GetCharconverts any letters received to upper
case, so that this expression will be True if the user types s, C, or q,
as well. Without this conversion, those letters would be ignored,
since lower case letters are different from upper case letters.
Besides testing for membership in a set, you can also make
comparisons between sets themselves. Two comparisons are
obvious: equality (=) and inequality « ».
The other two, "<="
and ">=", are for set inclusion. The expressions
Set! <= Set2
Set2 >= Set!
will return True if all the elements in Set1 are also in Set2.
Suppose you have the following:
16-3
type
ShipType = (Constltutlon,Enterpr1se,ReUant,Loknar,
Larson,Chandley,ExcelsIor,Baker)j
ShipSet = set of ShipTypej
var
AllShips,Frigates,Destroyers,Cruisers,Battleship,Temp
: ShipS etj
begin
AllShips ·= [Constltutlon..Baker] j
Frigates .= [Loknar,ChandleY]j
Destroyers .= [Larson,Baker]j
Cruisers ·= [Constltutlon..ReUant,Excelslor] j
Battleship ·= [ExcelsIor] j
Temp ·= [Constltutlon..Baker] j
end
Given this, the following expressions have the values shown:
AllShips = Temp True
AllShips = Cruisers False
Frigates < > Destroyers True
BattleShip <= Cruisers True
Temp >= AllSh1ps True
Destroyers < = AllShips True
Cruisers < = Battleship False
[] <= <any set> True
16.2 SET OPERATIONS
As has been mentioned, sets represent collections of values of a

scalar data type (I nteger, Byte, Boolean, Char, or a declared
scalar type). You've seen how to compare sets; you can also
perform several operations on them. Let's assume the following
definitions:
type
ShipType = (Constltutlon,Enterpr1se,ReUant,Loknar,
Larson,Chandley,ExcelsIor,Baker)j
ShipSet = set of ShipTypej
var
Set! ,Set2,Set3,AllSets,Empty
: Sh1pSetj
16-4
begin
.AllSets .= [Constitution..Baker];
Setl := [Constitution..Reliant,Excelsior];
Set2 := [Loknar,Larson,Chandley,Baker];
Set3 .= [Enterprise,Loknar,Baker];
Empty := [];
8D4.
The first operation to look at is set intersection, which uses the

multiplication symbol (*). The intersection of two sets is a set
containing all the elements common to both sets. For example,
here are some intersections, along with the resulting sets:
*
Setl Set2 []
*
Set2 Set3 [Loknar,Baker]
*
Set3 .AllShips [Enterprise,Loknar,Baker]
*
.AllShips Empty []
In each case, the result is a set containing the elements that
appear in both sets. In two cases (the first and the last), there were
no elements in common, so the result was the empty set.
The second operation is set union, which uses the plus sign (+).
The union oftwo sets is the set containing all the elements in both
sets. Here are some examples:
Setl + Set2 [Enterprise..Baker] ( = .AllShips)
Set2 + Set3 [Enterprise,Loknar,Larson,Chandley,Baker]
Set3 + Setl [Constitution..Loknar,Excelsior,Baker]
.AllShips + Empty [Enterprise..Baker]
The third and last operation is set difference, which uses the
minus sign (-). The difference of two sets is the set containing all
of the elements in the first set that are not in the second set. Note
well that, unlike set intersection and set union, the set difference
operation is not commutative; that is, Set1 -Set2 is not the same
as Set2 -Set1. Here are some examples:
16-5
AllSh1ps - Setl [Loknar..Chandley,Baker] (= Set2)

AllSh1ps - Set2 [Constltution..Reliant,Excelsior]
(= Setl)
Set! - AllSh1ps []
Setl - Set2 [Constitution..Reliant,Excelsior]
Setl - Set3 [Constltution,Reliant,Excelsior]
Set3 - Setl [Loknar,Baker]
Set3 - Set2 [EnterpriSe]
Set2 - Set3 [Larson,Chandley]
Note that if SetA - SetB = [ ] (the empty set), then SetA and
SetB are mutually exclusive, that is, they have no elements in
common.
16-6
Part 2 Chapter 17 Pointers
17. POINTERS AND

DYNAMIC ALLOCATION
-Polnters-
From time to time, you will find yourself wanting to be able to
create and destroy data structures while a program is actually
executing. Let's recall the Ship data type from the chapter on
records:
ccmst
XMax 8',
YMax 10;
type
XRange l..x:Max;
YRange l ..Y.Max;
Government (Federatlon,Kl1ngon,Romulan,Trader)i
Sides (Forward,FPort,FStarboardA}'ort,
AStarboa.rd,A.:tt);
Weapons array[Sld.es] of Byte;
17-1
Chapter 17 Pointers Part 2
AShip=
record
SX XRange;
BY YRange;
QX,QY Byte;
Energy 0 ..1023;
Federation (Phaser Weapons;
Photon array[Sides] of 0 .. 16);
(D1Bruptor Weapons;
Torpedo array[Sides] of 0 ..7);
(Beaml,Beam2 Weapons;
Plasma array[Sides] of 0 ..9;
Cloaked Boolean);
Trader ( )
end;
Now, let's suppose that the number of ships in existence during

the course of a game varies widely. What kind of data structure
would you use to be sure you could handle all (reasonable)
cases?
One possible solution, using arrays, might look like this:
var
Ships : array[I ..MaxShips] ofAShip;
ShipCount : O..MaxShips;
where MaxShips is a const value representing the maximum
number of ships allowed during the game. The variable Ship-
Count would be initially set to O. Each time a new ship was
created, the next free slot in Ships would be used:
if ShipCount < MaxShips then begin
ShipCount:= ShipCount + 1;
CreateShip(Ships[ShipCount])
end
else WriteLn('No more space for ships');
The procedure CreateShip is one that you would write. It would

set all of the fields of Ships [ShipCount] to the appropriate initial
values. When a ship was destroyed, say Ships[lndx], then you
would "shuffle down" all the ships from Indx+1 to ShipCount like
this:
17-2
for Jndx := Indx+ 1 to ShipCOWlt do

Ships[Jndx-1] := Ships[Jndx];
ShipCOWlt := ShipCOWlt - 1;
This method is fine, but has two drawbacks. First, you (the
programmer) must decide ahead of time the maximum number
of ships possible (MaxShips). The program will never be able to
handle any more ships than that. Second, space (memory) will
always be allocated for the maximum number of ships, regardless
of how many are actually in use. If you defi ne MaxShips to be 100,
then you can never have more than 100 ships, and space for 100
ships will always be set aside even if you only have one or
two. These are common limitations in most languages, but
Pascal does offer a way around them: pointers.
17.1 POINTERS
Suppose you modify the definition of Ships as follows:

var
Ships : array[l..MaxShips] of AAShip;
ShipCOWlt : O•.MaxShips;
If you'll look closely, you'll see that Ships is no longer an array of

AShip but of AAShip. The notation A on a data type refers to a
painter to the data type. I n other words, Ships is no longer 100
records (or whatever MaxShips is), but 100 pointers to records.
"What's the difference?", you may ask. Simple. The record is
some collection of data. The pointer is an address. If you still
don't see the distinction, consider this: what's the difference
between 100 people and the phone numbers of 100 people? With
either one, you can talk to 100 people, but you need an
auditorium for the former; for the latter, you just need a phone
and sufficient room for the people you choose to summon. The
analogy is a little rough, but you should get the idea.
The next question is, how do you use a pointer? Or, better put,
how do you use the record that the pointer points to? Answer:by
pointing to it. If Ships is an array of AAShip (that is, pointers to
records of. type AShip), then Ships[1] is a pointer-that is, an
17-3
address-and Ships[1]'" is the record to which it is pointing. In

other words, by sticking a carat ("') at the end of the pointer
variable, you now refer to the data structure to which it points. For
example, if you wanted to set the energy level of all currently
allocated ships to 1023, you would do the following:
for Indx :=Number to ShipCount do
Ships[Indx]"'.Energy:= 1023;
Note well that you cannot write Ships [lndx].Energy := 1023, since
Ships[lndx] is a pointer, not a record of type AShip. The carat
makes a big difference.
The third question is, how do you point the pointer? In other
words, how do you assign an address to the pointer, and how do
you ensure that the address the pOinter contains represents an
area of memory that is not being used by anything else (operating
system, program, other pointers, etc.)? Answer: Pascal does it for
you via the predeclared procedure New. For example, your ship
initialization routine could look like this:
if Sb1pCount < MaxSb1ps then begin
New(Ships[ShipCount]);
CreateShip(Ships[ShipCount]"')
end
else Y"frlteLn( 'Maximum number of ships aliocated');
Note that the procedure New takes a pointer as its argument. It
allocates the necessary amount of memory and sets the pointer
to the appropriate address. Also, note that there is a special
pointer value called nil which is used to indicated pointing at
nothing. Its special value is usable on all pointer types.
17.2 THE HEAP
The procedure New creates a new copy of the appropriate type of

data structure. For example, the statement
New(Ships[ShipCount]);
17-4
creates a new record of type AShip, the address of which

Ships[ShipCount) now contains. You may recall that a variable of
type AShip takes up 26 bytes. The question is, just where are
those 26 bytes? They're not in the array Ships, which just
contains pointers. Instead, they have been allocated in an area of
RAM known as the heap. Loosely put, the heap consists of
whatever memory is left over after allocating space for the
operating system (CP/M, CP/M-86, MS-DOS), TURBO Pascal (if
loaded), the run-time library, your program, and any variables
declared within the program (such as Ships).
If you're using an 8-bit system (with a maximum of 64K), have
TURBO Pascal loaded, and are running a large program, then the
heap may be very small; possibly as little as 1K, enough to hold
about 40 AShip records. On the other hand, if you're using a
16-bit system with lots of RAM, then the heap can be very big.
How big? A program running under MS-DOS 2.0 on a 512K
system has over 430K of memory in the heap, or enough space
for about 17,000 A Ship records. Because of those variations, and
because of the different operating systems involved, it's hard to
discuss specifics of how the heap works; that information is best
gleaned from the OS-specific appendices in the TURBO Pascal
Reference Manual.
Having read all that, you're probably still wondering how the
heap is used. When your program starts running, a special
variable (HeapPtr) points to the start (or bottom) of the heap.
When you call New(Ptr), then Ptr gets the value of HeapPtr, and
HeapPtr is increased by the size of the data structure that Ptr
points to. In other words, the sequence is something like this:
Ptr:= HeapPtr; HeapPtr:= HeapPtr + SizeOf(Ptr");
Repeated calls to New causes HeapPtr to "grow" upwards,

reducing the space remaining on t~e heap. You can find out at
any time just how much free space is left on the heap via the
standard function MemAvail. For 8-bit systems, MemAvail returns
17-5
the number of bytes left on the heap; for 16-bit systems, the
number of paragraphs (16-byte chunks). MemAvaii returns an
I nteger value, which means that it will be negative ifthere are over
32K bytes/paragraphs free. You can convert it to a positive
number with the following code:
var
TrueFree : Real.;
begin
TrueFree := MemAvail; { convert to real. value}
if TrueFree < 0.0
then TrueFree := TrueFree + 65536.0;
WriteLn('Space available: ',TrueFree:7:0)
end.
17.3 LINKED LISTS
You've partially solved your problem-you no longer have to

allocate space for the maximum number of ships-but you still
have to decide ahead of time what that maximum number is. And
since that means that you always have to allocate space for that
many pointers, you may not have gained that much of an
advantage. Is there no way around this problem?
Obviously there is, or else you wouldn't be reading this. The
answer lies in creating what is known as a linked list. Atthis point,
there is a temptation to refer you to Chapter 2 in the book
Fundamental Algorithms (2nd ed.) by Donald E. Knuth (Addison-
Wesley; Reading, Massachusetts; 1973) and leave it at that.
However, having brought the subject up, there results a certain
obligation to tell you something about it. Here goes.
Let's modify the definition of the record AShip as follows:
17-6
AShipptr = AAShip;
AShip=
record
SX XRange;
BY YRange;
QX,QY Byte;
Energy 0 ..1023;
ShieldB,Beams Byte;
Next AShipptr
end;
You may realize that you just have violated the Great Underlying
Rule of Pascal, namely that no identifier can be referenced until it
has been declared. As you can see, you defined AShipPtr to be a
pointer to type AShip before you defined AShip. This ability is a
necessary exception, since without it linked lists would be
impossible (or, at least, very messy).
Having modified your data structure, you now define your
variables as follows:
var
F1rst,Last : AShipptr
ShipCount : Integer;
At the start of the game, you would initialize your variables as
follows:
First := nil;
Last: = nil;
ShipCount := 0;
(The predefined identifier nil represents the value a pointer has
when it's not pointing at anything.)When you wish to create a
new ship, you call your procedure AddShip:
17-7
procedure AdclShip;
begin
if First = Dil then begin
New(First);
Last := First
end.
el8ebegin
New(L8Bt" .Next);
Last := Last" .Next
end;
Last" .Next := Dil;
Creata<3hip(L8B1i' )
end; { of proo AddShip }
(Note that you have a special test for creating a ship when none
exist.) The pointer First always points to the first ship in the list.
Each ship then points to the next one in the list via the field Next.
, The pOinter Last always points to the last ship in the list. If there is
only one ship in the list, then First and Last both point to it. If there
are no ships at all in the list (the list is empty), then both First and
Last equal nil.
Now that you can create the list, how do you reference a
particular ship in it? With the array it was easy: all you had to do
was index into the array and hey, presto! there it was. With the
linked list, you have to look for it. Here's a procedure to do just
that. It takes the index value and returns a pointer to the
appropriate ship (if it exists):
procedure FetohShip(Index : Integer;
var Ptr : .ABhipPtr);
begin
Ptr := First; { point at first ship in list }
while (Index: > 0) 8D4 (Ptr < > Dil) do begin
Index: := Index: - 1;
Ptr :=Ptl"".Next
end.
end; { of proo FetohShip }
If Index is less than or equal to 0, then Ptr points to the first ship in
the list. If Index is greater than the number of ships in the list, then
Ptr gets set to nil. Otherwise, Ptr points to the appropriate ship.
17-8
17.4 DEALLOCATION AND MEMORY MANAGEMENT
At this point, you need to know about how to reclaim the memory
used by a pointer when you no longer need the data structure it
points at. TURBO Pascal provides two approaches to memory
management. One method is to use the predefined procedure
Dispose to reclaim the memory pointed to by a given pointer.
Going back to your array of pointers (Ships), you could delete a
particular ship this way:
ifShips[Indx] < > nil
then Dispose(Ships[Indx]);
This would set Ships [Indx] equal to nil and makethe memory that
it had used available for other pointers. If you just wrote
Ships[Indx] := nil;
then the memory would not be reclaimed because the pointer

would simply be pointing at nil.
Even when memory is reclaimed (via Dispose), there arise the
problems of memory management, fragmentation, and garbage
collection (honest, that's what it's called!). Without going into all
of the gory details (again, see Knuth), here's a brief, simple
description of the problem.
Think ofthe heap (the memory used for dynamic allocation) as a
long, narrow shelf. Each time you create a variable using New,
you place a wooden block (width = amount of memory needed)
somewhere on the shelf where it will fit. Some decision has to be
made about where to place it. That's memory management. Each
time you destroy a variable using Dispose, you remove the
corresponding block from the shelf. After repeated calls to New
and Dispose, you can find that the free space on the shelf is tied
up in lots of small, often useless chunks between blocks. That's
fragmentation. To reclaim all of those little chunks, you shove all
of the wooden blocks together and move them to one end of the
shelf. Now, all of the free space is in one large chunk. That's
garbage collection. It all sounds easy, but in practice it can be a
real headache.
17-9
TURBO Pascal supports New, Dispose, and MaxAvail (which

returns the size of the largest tree block of memory on the heap),
but it does not have a garbage collection facility. If, however, all
the data structu res you create and destroy are the same type, or
even just the same size, then fragmentation and garbage col-
lection are not problems. Why? Because any gap that appears
because of a given item being disposed will be exactly the right
size for any new item being created. In other words, if all the
wooden blocks on the shelf are the same size, then you'll never
have a problem adding and deleting them (except, of course,
when you run out ot space altogether).
If you foresee problems using Dispose, you might consider
TURBO's alternate memory management scheme, a very simple
one. When you create variables using New, the blocks are placed
next to each other, starting at the left end of the shelf (low
memory). Atany pOint, you can usethe procedure Markto get the
address where memory for the next variable would be allocated.
You can then continue to create variables. Later, you can call the
procedure Release with the value you received trom Mark. This
effectively clears everything off the shelf that's to the right
(above) the address passed to Release. Variables now created
will be allocated starting at the address read when you called Mark.
There are some important things to note here. Let's suppose that
you executed the following code:
First: = nil;
Sh1pCount := 0;
Mark(TempPtr );
AddSh1p;
Release(TempPtr);
You have created 10 ships, then reset your next-tree-location
pOinter (called the heap pointer) to where it was before those ten
ships were created. This has done nothing to affect the linked list of
ships. All the information is still there, all ofthe pointers are still cor-
rect. The call to Release has no effect until the next time you call
New. At that time, the newly-created data structure would start to
overwrite the linked list. It you then read the linked list, tunny
things could result. The moral? If you use Mark and Release, be
sure to set to nil any pointers pointing into the treed-up memory.
17-10
Part 2 Chapter 18 Files
18. FILES
Throughout this book, you've probably noticed the heavy

emphasis Pascal places on data structures: scalar variables,
declared scalar types (DSTs), records, arrays, sets, and the like. It
is no coincidence that when Niklaus Wirth, the father of Pascal,
wrote a book on software development, he entitled it Algorithms =
Data Structures + Programs. You've seen all kinds of data
structures in the examples in this book, and you've been shown
just how powerful they can be.
But, as always, a new problem has cropped up. Sure, you can
define packed records with variant tag fields, or you can create
arrays of sets of DSTs. But what happens when the program
ends, as it sooner or later must? All your data structures quietly
vanish away, and the information they hold is lost. Of course, you
often don't care; when the program is done, you're done, too, and
you no longer need the information. There are times, however,
when you would like to save that data for the next time you run the
program. For example, if you are playing a game, you may want
to (or have to) stop for a while (hours, days, weeks). It could be
disastrous to have to start over again from scratch.
There is, of course, a solution to this problem: files. A file is a
collection of data structures, all ofthe same type. It is similar to an
array, but with one big difference: most of the information is out
on a disk. If that doesn't tell you much, don't worry. The following
example should help clear things up:
18-1
Chapter 18 Files Part 2
cOD.St
ShipMax 40;
XMax 8;
YMax 10;
tine
XRange 1 ..XMax;
YRange l ..YMax;
AShip
record
SX XRange;
BY YRange;
QX,QY Byte;
Energy 0 .. 1023;
ShieldB,Beams 0 ..266
end;
var
ShipFtle file of AShip;
Ship array[ l ..ShipMax] of AShip;
ShipCnt O..ShipMax;
The array Ship is a collection of 40 records of type AShip. All of

those records are always in memory, and anyone ofthem can be
referred to using the n"otation Ship [Indx] , where Indx is an integer
variable or expression resolving to a value from 1 to 40 (ShipMax).
All the information in Ship is lost when the program is done.
The file ShipFile manages a collection of some number of
records of type AShip. All of those records are kept out on mass
storage (floppy disk, hard disk, RAM disk, whatever). ShipFile
represent a "door" through which you can get or put individual
records. However, it does it in a manner somewhat different from
Standard Pascal. Standard Pascal uses the file variable (ShipFile)
as a poi nter; you then use the standard proced u res Put and Get to
copy records to and from the file. TURBO Pascal simply uses the
Read and Write statements, along the same lines as textfile 1/0.
The best way, of course, to show you what this all means is to give
an example. Let's suppose that you've been playing your game
for a while, and you want to quit, but you want to be able to
resume your game later. Among other things, you need to save
on the disk the information in Ship. The variable ShipCnt
indicates how many ofthe records in Ship you are actually using.
You then could write those records out to disk as follows:
18-2
Ass1gn.(ShipFlle,'SHIPS.DAT'); { ass1gn. to a file on diSk }

Rewrlte(ShlpFlle); { open the file for writing}
for Indx := 1 to ShlpCnt do
Write(ShlpFlle,Shlp[Indx]); { put the next record out}
Close(ShipFlle); { close and save the file}
The Assign statement connects the file variable ShipFile to the
name "SHIPS.OAT". No disk file has actually been accessed at
this point. The Rewrite statement, however, does cause some
action. A file named SHIPS.OAT is now created outon the disk; if
a file by that name already exists, it is erased. The file variable
points at the first (Oth) record in the file; the file itself is empty.
Each call to Write copies the record contained in Ship [Indx] out
to the disk file and advances to pointer to the next position. Close
makes sure that any records being held in memory by the
operating system are actually written out to the disk. and closes
out the file, breaking the connection between ShipFile and
"SHIPS.DAT".
You've saved the ship information out to disk and have stopped
playing. Now, after a while, you want to start the game up again
where you left off. Among other things, you need to read the
records you saved back into Ship. One problem faces you,
though: how does the program know how many records there are
in the file? Answer: it doesn't need to. Instead, Pascal provides
the boolean function EOF. EOF stands for end of file. As you read
the file, EOF returns a False value until you hit the-end; then it
returns True. You can now read in the records as follows:
Ass1gn(ShlpFlle,'SHIPS.DAT'); { assign to a diSk file}
Reset(ShlpFlle); { open the file for reading}
ShipCnt := 0; { # of shlps in arraor}
while DOt EOF( ShlpFlle) do begin
ShlpCnt:= ShlpCnt + 1; {yes, there is another ship}
Read(ShlpFlle,8hlp[ShlpCnt]);
end;
Close(ShlpFlle)j { close file}
The Assign statement works as before. This time, though, you
use Reset to start reading from the file. If SHIPS.OAT doesn't
exist, you'll get an I/O error (see the section on Error Handling
later on); otherwise, the file variable will point at the first (Oth)
record. Each call to Read copies the record currently pointed at
18-3
into Ship [ShipCnt] and then points at the next record. If there is
no next record, then the function EOF starts returning True, and
you drop out of the end of the loop. Since you didn't know ahead
of time how many records there were, you had to count them (by
incrementing ShipCnt) until you reached the end of the file.
Actually, with TURBO Pascal, you could have found out ahead of
time just how many records there were (though it really wouldn't
matter). The function FileSize will return the number of records in
a file, so you could have written:
ABs1gn(ShipFlle,'SHIPS.DAT');
Reset( ShipFlle)j
ShipCnt:= FllaSize(ShipFlle)j
for Indx := 1 to ShipCnt 40
Read(ShipFile,8hip[Indx] );
Close(ShipFile) j
Both approaches work equally well; the second one has the
advantage of letting you know ahead of time just how many
records you have to read in (in case that's important).
In this example, ShipFile was a file of records of type AShip. You
can define a file of anything (except files), so you could have
done this instead:
type
AShip =
record
SX XRangej
BY YRangej
QX,QY Byte;
Energy 0 .. 1023j
Beams,8hieldB 0 ..266
end;
ShipArraor = array[l ..ShipMax] of AS hip;
var
Ship ShipArraor;
ShipFile file of ShipArraorj
ShipCnt O..8hipMaxj
18-4
Now, to save the information in Ship, you could write:

Assign(ShipFlle,'SHIPS.DAT');
Rewrite(ShipFlle);
Write( ShipFlle,Ship);
Close(ShipFlle);
Simple, isn't it? With a single Write statement, you have sent all
the ship records out to the disk. To read it back in, you just write:
Assign(ShipFlle,'SHIPS.DAT');
ResetCShipFile);
Read(Ship:B'ile,Ship);
Close(ShipFile);
Unfortunately, you've now lost track of one piece of information,

namely the value of ShipCnt. However, there are a number of
ways around this. First, you could redefine ShipArray to be a
record with two fields: the ship count and the array itself. Second,
you could save the value of ShipCnt in another file, either by itself
or as part of another data structure. Third, you could add another
field to the definition of AShip which would indicate whether or
not that record was, indeed, a "real" ship; then, after you've read
Ship in, you just go through and count all of the records with that
value set. Finally, you could rewrite the game so that the value
ShipCnt is no longer needed.
18.1 RANDOM ACCESS OF FILES

From what you've seen so far, to read or write a particular record
in a file, you must go to the start of the file and work from there. If
you wish to read the 10th record in a file, you must reset the file
and go through the first nine records to get to the tenth. If you
want to modify the tenth record, you must read it in as mentioned,
reset the file, read in the first nine, then write the 10th back out. In
short, a lot of work.
18-5
Now let's suppose that you want to sort that file in some way.
Your sorting program would have to do a lot of reading from and
writing to the file, using the clumsy procedure above. If the file is
small enough, you can read the entire thing into memory, sort it
there, then write it back out. If it isn't, you're in for a long, tedious
program.
Your task would be easier if you could read and write any given
record in the file without having to go back to the start and read all
the preceding records. Such a capability is known as random
access of files, as opposed to the sequential access described
above. Since this is such a desirable trait, TURBO Pascal offers it
to you via the Seek statement, taking the form
Seek(FlleVar,RecNum);
where FileVar is the file variable and ReeNum is an integer

variable or expression indicating the record number. After
executing such a statement, Read{FileVar,Ree) will copy record
#ReeNum into Ree, while Write{FileVar,Ree) will write the data
inRee out to record number #ReeNum in the file. Since (as
mentioned above) the first record is record #0, the statement
Seek(FlleVar,O);
will pOint to the first record in the file. By the same token, the
statement
Seek(FlleVar,Flla91ze(FlleVar)-1);
will point at the last record in the file. Seek can, in fact, be used to
expand a file, that is, to add records onto the end of it. The
statements
Seek(FlleVa.r,Fila91ze(FlleVar));
Write(FlleVar,Rec);
will append Ree onto the end of the disk file that FileVar is
assigned to. You can continue to do this until you run out of disk
space. Note, though, that you cannot seek beyond FileSize{File
Var); that, too, will result in an I/O error.
18-6
There is one more function that TURBO Pascal provides for

random access of disk files: FilePos. The expression FilePos(File
Var) returns the number (O .. FileSize) of the record that FileVar is
currently pointing at.
18.2 TEXT FILES

Pascal differs from BASIC, FORTRAN, and several other lan-
guages in that it writes data files as binary files, rather than text
files. In other words, Pascal stores a two-byte integer value as a
two-byte integer value, not as a character string several bytes
long, with spaces and carriage returns as delimiters. This results
in smaller data files and faster disk I/O (since there is no
conversion from data to ASCII and back). This also allows you to
do things like read in or write out an entire array with a single
command, rather than having to use intricate loops and heavily
formatted I/O statements.
Often, though, you do need to read and write text files-for
example, when you are reading in user-entered data or writing
out reports or other information. Pascal allows for that, too. You
can use the predefined data type Text, which is equivalent to file
of Char. (The latter is not available in TURBO Pascal.) So, for
example, you could define:
program. FlleTest;
var
InFlle,OutFlle : Text;
begin
Ass1gn(InFlle,'INPUT.TXT')i
Reset(InFlle)i
Ass1gn(OutFlle,'OUTPOT.TXT');
Rewrlte(OutFlle)i
8D4. { of program FlleTest }
You can now read in text from InFile, and write it out to OutFile,
using the following procedures:
18-7
Read(F,Pl, ... ,Pn) reads in the variables PI through

Pn from text file F
Readln(F,Pl, ... ,Pn) reads a.s above, plus advances to
next line
Write(F,Pl, ... ,Pn) writes variables PI through Pn into
text file F, all on same line
WriteLn(F,Pl, .. ,Pn) writes a.s above and advances to next line
These routines are pretty much what they appear to be. The
parameters can be variables (and, for Write and WriteLn, con-
stants and expressions) of type Integer, Real, Char, and String
(Write/WriteLn can also handle Boolean). ReadLn and WriteLn
don't need to have parameters at all; they can simply be used to
read and write end-of-line markers. If you leave the file variable
(F) out, then Read and ReadLn default to keyboard input, while
Write and WriteLn default to screen output.
Let's say that you wanted to write out the data for your game in
textfile format, and had InFile and OutFile declared as Text. Our
game save routine might look like this:
Assign(OutFile,'SHIPS.TXT')j
Rewrite( OutFile);
Wr1teLn(OutFile,8h1pCnt:4 )j
for Indx := 1 to sh1pcnt do
with Sh1p[Indx] do begin
Write(OutFlle,SX:4,BY:4,QX:4,QY:4)j
WriteLn(OutFlle,Energy:5,Beams:4,8h1elds:4)
end;
Close( OutFlle);
(The formatting notation, p:n, means to right-justify the value ofp

within a field of n characters. More on this later.)
Your ship information has now been stored out as a text file, one
which you could edit or print. If you had three ships, at various
locations, with full energy (1023), and with beams and shields set
at 0, the file SHIPS.TXT might look like this:
3
7 2 50 175 1023 o o
3 10 199 12 1023 o o
5 5 o 47 1023 o o
18-8
Having saved your game information as text, you can (and must)
read it back in the same way. The following code would suffice:
ABs1gn(InFtle,'SHIPS.TXT');
Reset(InFtle);
Rea.dln(InFtle,8hipCnt);
far Indx := 1 to ShipCnt do
with Ship[Indx] do
ReadLn(InFtle,8X,8Y,QX,QY,Energy,Beams,8hields);
Close(InFtle);
18.3 THE EOLN FUNCTION

Earlier, we discussed the Boolean function eof(f) which tells you
when you've reached the end of a file. However, text files have an
additional delimiter: end of line. At times, you may want to read in
a line item by item and know when you've hit the end of the line.
How can you tell? By using the Boolean function eoln(f), which
tells you just that.
Let's suppose that you have the following text file:
342 3 1123
41 1772 1 2 33 6
66
441 233 23 67
You don't know ahead oftime how many numbers there are on a
line, but you still need to read them in. The following chunk of
code will read them into an array:
program eolntest;
CODSt
nummax: = 60;
var
1nfile text;
numlist array[ 1 ..nummax] of integer;
numcnt integer;
18-9
begin
reset(1nflle); { agam, this will vary }
numcnt := 0; { noth1ng in arrB8 yet}
while DOt eott 1nflle) aDd. (numcnt < numm.a.x) do begin
whue DOt eoln( 1nflle) aDd. (numcnt < nummax) do begin
numcnt := numcnt + 1;
read( 1nfile,num11st[ numcnt])
end.;
readln(1nfile); { read in end of line}
end.;
for inch( := 1 to numcnt do
wrlteln('num11st[',indx:2,'] = ',num11st[indx]:6); close(1nfile)
end. { of program eolntest }
In the innermost loop, you read numbers off of a line, one by one,
until you hit the end of the line (eoln). You then advance to the
next line (using readln) and continue the process until you either
hit the end of the file (eof) or have filled up the array (numcnt =
nummax). At the end, for a check, you write all of the values out to
the screen.
18.4 . FORMATTED OUTPUT
Pascal has been accused of having lousy I/O capabilities, usually

by peoplewho are familiarwith BASIC, FORTRAN, etc., and who
are used to the extensive I/O formatting commands available.
Actually, TURBO Pascal has extremely powerful 1/0 capabilities
and is capable of doing things that cannot even be done if you're
programming in BASIC or FORTRAN.
TURBO Pascal contains formatting capabilities. You can write
out Integer (and Byte), Real, Char, Boolean, and String values
(variables, constants, expressions). For all of those, you can
specify a field width, using the notation Parm:N, where N is the
width (in characters) of the field. If N is greater than the number of
characters that Parm actually requires, then blanks are added in
front of Parm to expand it to the proper width (this is called
right-Justification, because you're forcing Parm to be flush with-
or justified against-the right-hand side ofthe field). For example,
the statements
18-10
Indx:= 16421;
WrlteLn('This field is too narrow:' ,Indx:3);
WrlteLn( 'This field is just right:' ,Indx:6)j
WrlteLn('This field is too wide :' ,Indx:16);
would produce
This field is too narrow: 16421
This field isjust right: 16421
This field is too wide : 16421
Note that if you specify a field that is too narrow, the field
automatically expands by the minimum amount necessary to
write the value out. Specifying no field at all is equivalent to giving
one that is too short: just enough space is allocated to write the
value out, and no more. This can cause problems if you're writing
more than one value per line. For example, if you had written
Wrlte(OutFUe,SX,SY,QX,QY);
WrlteLn( OutFUe,Energy,Beams,8hieldB);
in the code to save the ship information out to a text file, the file
SHIPS.TXT would look like this instead:
3
7260176102300
31019912102300
66047102300
Reading the information back in would be impossible, since you
would be unable to separate the fields into their different values.
So it pays to use a width specifier whenever appropriate.
As mentioned above, you can write out Boolean values, with or
without a width parameter. The values translate into the strings
'TRUE' and 'FALSE', so that the statements:
Flag1 := Truej Flag2 := Falsej
Wrlteln('Flag1: ' ,Flag1:6,' Flag2: ' ,Flag2:6);
would produce the output
Flag1: TRUE Flag2: FAIBE
18-11
You cannot, however, read in TRUE and FALSE as boolean

values. Instead, you would have to read them in as strings and
convert accordingly.
Real values have their own special output rules for TURBO
Pascal. If no field width is specified, then the value is written out in
an 18-character-wide field, with the format
bsdddddddddddEtdd where b = blank (' ')
s = blank or '-'
d = digit ('0' ..'9')
t = '+' or '-'
If a single field width (Parm:N) is given, then the same basic

format is used, with some adjustments for field width:
<b>sd.<d>Etdd where <b> is 0 or more blanks
<d> is 1 to 10 digits
The minimum value for N is 7 (8, if Parm < 0.0). As N gets large,
the number of digits is<d> expands until it hits 10. After that, the
number of blanks in <b> starts increasing. The format given
above for no width parameter is equivalent to Parm:18.
You can specify a second field width for reals (Parm:N:M). In that
case, M represents the number of digits after the decimal point
that should be displayed, and fixed-point (rather than floating
point) notation is used. M must be less than or equal to N-3. If M
equals 0, then no decimal point and no following digits are
displayed. If M is greater than 24, then floating-point notation is
used.
If all of this seems confusing, here are some examples to show
you how everything works:
18-12
program FormatDemo;
const
P1 = 3.1416926636;
begin
Wr1teLn(P1);
Wr1teLn(P1:8);
Wr1teLn( -P1:8);
Wr1teLn(P1:12);
WrlteLn(P1: 16);
Wr1teLn(P1:20);
Wr1teLn(P1:8:0);
Wr1teLn(P1:8:4);
WrlteLn(P1:12:10)
end. { of program FormatDemo }
will produce the output

3.1415926535E+OO
3.14E+OO
-3.1E+OO
3.141693E+OO
3.1416926636E+OO
3.1416926636E+OO
3
3.1416
3.1416926636
18.5 FILENAMES
TURBO Pascal runs under CP/M, CP/M-86 and MS/OOS (a
version is available for each), so it follows the file conventions of
each system, which are essentially the same: a file name of up to
eight (8) letters and/or digits, with an optional 3-character
extension. If the extension exists, it is separated from the file
name by a period. For example, the following file names are
acceptable:
th1sflle
thx1138
Simple.pas
Simple. axe
stars.dat
18-13
If you use these file names "as-is", then your program will
assume that they are on the currently logged drive (as defined
either by TURBO Pascal, if it's loaded, or by the operating system
prompt "X>", where "X" is the logged drive). You can, of course,
explicitly state which drive the file is on by appending the drive
name in front of the file name:
a.:th1sflle
b:thxl138
c:stars.dat
18.6 UNTYPED FILES

Up until now, all files you've looked at have been textfiles (= Text)
or data files (= file of <type». However, there are times when you
want to deal with "raw" data, i.e., data that hasn't been formatted
for you. For example, in some programs, you need to have many
different types of data out on the disk. However, you don't want
the overhead of having many files open at once or having to
constantly open and close files.
The solution? You put everything into one large disk file and use
an untyped file to access it. An untyped file is declared as follows:
var
B1gFlle : IDe;
It's opened and closed just like any other file. Reading from and
writing to it are different, though, from other files. Untyped files
can only be read or written a block ("128 bytes) at a time. You do
this using two procedures, BlockRead and BlockWrite. Here's
their format:
BlockRead(FlleVar,Buf,NumBlocks);
BlockWrlte(FlleVar,Buf,NumBlocks);
Fi/eVar is, of course, the untyped file variable (such as BigFi/e

above). Buf can be any type of variable; most often, it's an array of
some sort. NumBlocks is the number of blocks that you want to
read or write. Since each block is 128 bytes, SizeOf(Buf) must be
18-14
greater than or equal to 128*NumBlocks. This is critical. These

two routines do not do any range-checking, and if Buf is too
small, then a BlockRead will cheerfully copy the requested
blocks to memory, overwriting whatever code and/or variables
follow Buf. A BlockWrite in such a case is not necessarily as
disastrous-you merely copy extra garbage out to the file-but
even that can catch up with you later.
All of the standard file routines (except for Read, Write, and
Flush) work on an untyped file. This means, among other things,
that you can do random access on the file. For an untyped file,
Seek will assume that each block is one record, with the first
block being record #0. You can directly move to a given block,
then use BlockRead and BlockWrite to access it (and those that
follow it). A word of caution, though: if you are opening an
existing file to update it (with BlockRead and/or Block Write) , be
sure to use the Reset procedure. Just as with data files and text
files, a call to Rewrite erases any existing file and creates a new
one.
Let's illustrate all this with an example. Suppose that our "Trek"
program has one data file to contain all game information:
sectors, ships, and so on. Suppose that blocks 10 through 13 of
your data file contain records which, for convenience's sake, are
32 bytes long each. Each block then has 4 of these records in it,
and there are 16 (4 * 4) records in all. You could then write the
following routine to get or put a specific record (numbered 0
through 15) from that chunk of the file. Let's say that record type
is BaseRec and that your untyped file (BigFile) is already open.
procedure BaseRecIO(Indx : Integer;
ReadFlag : Boolean;
var Ree : BaseRee);
{
does read/write for Ree[Indx]
ifReadFlag = True, then reads Ree, else writes it
}
var
Bent)B~ee)Err : Integer;
: array[O ..3] of BaseRee;
18-15
begin
Indx := Abs(Indx) mod 16; { force to allowable range}
IBlk := Indx div4 + 10; { calculate block #}
IRec := Indx mod 4; { calculate rec w lin block}
Seek(BtgFlle,IBlk)
BlockRead(BtgFlle,Data,IBlk) { get the data }
if ReadFlag
thenRec:= Data[IRec] { get appropriate record}
e1sebegin
Data[IRec] := Rec; { else save it in 'data'}
Seek(BtgFlle,IBlk)
BlockWrite(BtgFlle,Data,IBlk) {&write it}
end
end; { of proc BaseRecIO }
By writing similar routines for the other data types stored in

BigFile, you can have ready access (and random access, at that)
for a wide variety of data types with a minimum of overhead. If the
size of the buffer (128 bytes or some multiple thereof) bothers
you, you can always use New to create it on the heap and Dispose
(or Mark and Release) to get rid of it. ActuallY,since it is a local
variable, it is already allocated and deallocated on the fly.
18.7 DEVICE I/O
.~ JlnputlOutput-
18-16
Most applications require I/O involving the computer hardware

itself. Reading from the keyboard and writing out to the screen
are the two most obvious examples. Standard Pascal (and
TURBO Pascal) predefines the textfiles Input and Output for just
those functions. All Read and Write statements without a file
variable use these two files. But there are other times when you
might want to read from or write to a specific device. How do you
do this?
Simple. You use a set of special filenames that refer to hardware
devices rather than to disk files. Consider, for example, the
following program:
procedure Convert;
ccmst
OutFlleNarne = 'STABS.DAT';
type
StarArray = array[ 1..3] of Real;
var {Stars[l] = X, Stars[2] = Y, Stars[3] = Z}
InFlleName : striDg[30];
InFlle : Text;
OutFlle : file of StarArray;
Star : StarArray;
begin
Write('Enter input flle: '); ReadLn(InFlleNarne);
Assign(InFlle,InFlleN arne);
Reset(InFlle);
Assign( OutFlle,OutFlleNarne);
Rewrite( OutFlle);
while not EOF(InFlle) do begin
ReadLn(InFlle,star[ 1] ,star [ 2] ,star[ 3]);
Write( OutFlle,star)
end;
Close(InFlle); Close( OutFlle)
eD4. { of program Convert }
This program prompts the user for the name of an input file,
which it expects to be a text file with three real numbers on each
line. It reads those three numbers into Star, then writes Star out to
the data file STARS. OAT. In short, it's converting the star
information from a text file to a data file.
18-17
Now, suppose you wanted to enter the data manually instead of

having it read from a disk file. All you would have to do is give the
filename 'CON:' when asked for the input file. You would then
enter the data, line by line. When you weredone, you would type
ctrl-Z, which tells the program that it's reached the "end of file".
By the same token, a program that writes a text file to disk could
be redirect to write the output to the screen by also using 'CON:'.
CON: is known as a logical device, and there are several other
logical devices as well, representing keyboard input, screen
output, the printer, the serial port, and other such items. Note that
because these are character-oriented devices, only files of type
Text should be connected with them.
Here's a list of the special filenames that TURBO Pascal

recognizes:
CON: console I/O, i.e., read from the keyboard and write
to the screen. Echoes input, allows correction with
backspace. Expands tabs on output. Echoes CR
(carriage return) as CR/LF (carriage return/line
feed) for both input and output.
KBD: keyboard input with no echo or interpretation.

TRM: console output with no interpretation.
LST: the line printer. Can only be used for output. No
interpretation. Tabs are not expanded.
AUX: input and output device, usually an RS232 port.
Corresponds to PUN: and RDR: in CP/M.
USR: user I/O device. Advanced programmers can write
their own I/O drivers for specific devices.
In Convert, you wanted to be able to specify the input file when

you actually ran the program. Sometimes, you'll know ahead of
time that you want to do I/O with a specific device. In that case,
you don't have to define the file variable, assign it, and so on;
instead, TURBO Pascal has predefined (and pre-assigned) file
variables for each of these devices. In other words, you can use
the file variables Com, Trm, Kbd, Lst, Aux, and Usrwithout having
to do anything.
18-18
18.8 "REAL-TIME" KEYBOARD INPUT
I n certain applications (such as games), you want the program to

continue to run while at the same time checking for user input. In
other words, the program doesn't sit and wait for you to enter a
command, but instead would periodically check for and handle
user commands. For example, the "Trek" game would be
considerably more exciting if the attacking ships continued to
advance and fire while you were trying to make up your mind
what to do.
TURBO Pascal makes this fairly easy, via the Kbd device and the
Boolean function KeyPressed. As you might guess, KeyPressed
returns True if the user has hit any key, and False otherwise. It
does not sit and wait, but merely checks the console status. You
might, then, write a routine like this:
procedure CheckColIl.lIlB.n.d;
var
Cmd : Char;
begin
:if KeyPressed then begin
Read(Kbd,Cmd); { read key w lout echo}
Cmd:= UpCase(Cmd); { force to upper case}
caseCmdof
{handle commands}
else
Write(Chr(7)); {beep at illegal. cmd}
end
end
end; { of proc CheckCommand }
By scattering calls to CheckCommand through your program,

you can give the illusion of "real-time" commands, with swift
response to any user input at any point. You do need to be careful
of one thing, though: make sure that you don't call Check-
Command at a point where any of the commands could have
18-19
side-effects on what's currently being done. For example, you

shouldn't call it if the program is updating the display and
CheckCommand can change that display. Instead, call Check-
Command just before or just after that section of code.
18.9 1/0 ERROR HANDLING
Suppose that you ran Convert (the program listed above), and
that you gave it the name of a file that didn't exist. What would
happen? As soon as the program tried to open that file for input
(via the Reset call), you would get an I/O error, and the program
would abort. This, of course, is something of a pain, more so if
you were in the middle of a large, complex program rather than
just at the start of a small, simple one. Fortunately, TURBO Pascal
offers a solution to this problem. Using a compiler directive, you
can disable the "abort on I/O error" feature for sections of code
(or, if you wish, for the entire program). To turn off I/O error
trapping, you insert the comment statement { $I-} into your
program. When you want to turn it back on, you insert { $I+}. For
example, you could rewrite part of Convert to read
{ $I-}
Write(/Enter input me: '); ReadLn(InFileName);
ABsign(InFile,InFileName);
Reset(InFile);
{ $I+}
Of course, this by itself doesn't really solve your problem; the

program won't bomb, but if the file doesn't exist, you won't be
able to read in any stars. TURBO Pascal again comes to the
rescue with a bUilt-in function named IOresult. IOresult returns
an error code based on the success or failure of the last I/O
operation you tried to perform. If the operation was successful, it
returns the value 0; otherwise, it returns a value indicating just
what the problem was. For example, we could again rewrite the
code above like this:
18-20
{ $l-}
repeat
Wrlte('Enter input file: '); ReadLnClnFlleName);
Assign(lnFlle,lnFlleName);
Reset(lnFlle)
untU lOresult = 0;
{ $l+}
The liD compiler directive, along with IOresult, can be used in

many places to develop "bulletproof" programs, that is, programs
which cannot be bombed through various problems.
There are some important things you need to be aware of in using
these techniques. First, any call to IOresult returns the current
value, then resets the error condition to O. This means that you
can't keep calling IOresult and expect it to continue to return the
error code for the last problem you ran into. Suppose you had
written
{ $l-}
repeat
Wrlte('Enter input file: '); ReadLnClnFlleName)j
Assign(lnFlle,lnFlleName);
Reset(lnFlle)j
if lOresult > 0
then WrlteLnC'Flle not found; try again')
untU lOresult = OJ
{ $l+}
If you entered the name of a non-existent file, you would get the
message 'File not found; try again'. However, the call to IOresult
that caused that message would clear the error condition.
IOresult would then always return 0 at the until statement, and the
program would never go back and ask you to re-enter the file
name. What you need, instead, is something like this:
{ $l-}
re»eat .
Wrlte('Enter input file: ')j ReadLnClnFlleName)j
As sign(lnFlle,lnFlleName)j
Reset( lnFlle)j
lOerr := (lOresult < > O)j
if lOerr
then WrlteLnC'Flle not found; try again')
untU not lOerr;
{ $l+}
18-21
where 10err is a variable of type Boolean. That way, 10err can

"remember" that there was an I/O error, even after 10resuit has
been reset to O.
There is a second caution in using { $I-} / { $I+} and 10result. If an
error does occur, your program must call 10result before
attempting addition I/O. In other words, if you attempt to do I/O
when the error code is something other than 0, problems can
occur (such as your program hanging). Because of this, it is a
good idea either to limit your use of { $1- }If $1+ } to very specific
sections of your program, or to write your own I/O checking
routine and then call that anywhere where problems might occur.
Assuming that you have set aside line 24 of the screen for error
messages, your routine might look like this:
program MyProgram;
cODBt
Bell : Char = Chr(7);
10 err : Boolean = False;
{ These are both typed constants }
type
Prompt : BtriDg[ 80];
procedureError(Msg: Prompt);
{
write error Msg out on line 24 of the screen
}
begin
GoToXY(l,24); ClrEol;
Write(Bell,Msg)
end; { of proc Error}
18-22
l)rOC8Clure IOcheclq
{
check for I/O error; print message if needed
}
var
10 code :Intege~
Ch : Char;
begin
IOcode:= IOresult;
IOerr := (IOcode < > 0);
if IOerr then begin
case IOcode of
$01 : Error('File does not ex1st');
$02 : Error('File not open for input');
$03 : Error('File not open for output');
$04 : Error('File not open');
$10 : Error('Error in numeric format');
$20 : Error('Operat1on not allowed on logical dev1ce');
$21 : Error('Not allowed in direct mode');
$22 : Error('Assign to sta.nda.rdfUes not allowed.');
$90 : Error('Record length m1smatch');
$91 : Error('Seek beyond end-of-fUe');
$99 : Error('Unexpected end-of-fUe');
$FO : Error( 'Disk write error');
$Fl : Error('D1rectory is full');
$F2 : Error('File size over:tlow');
$FF : Error('File disappeared')
else
Error('Unlmown I/O error: ');
Wr1te(IOcode:3)
end; { of case }
Read(Kbd,Ch)
end.
end; { of proc IOcheck }
This routine does a number of things. First, since it always calls

10result, it clears the pending I/O error code. Second, it sets the
global flag 10err, so that other parts of the program will know
whether or not there has been an error and can act accordingly.
Third, it prints out an error message on line 24, pausing until the
user hits any key (hence "Read(Kbd,Ch)", which will read a
single character without echoing it back to the screen). Fourth, it
uses the else clause of the case statement to handle any
undefined I/O errors.
18-23
(A programming note: this routine uses a single case statement

for clarity. However, since the values in the case statement are so
widely spread-using only 15 values within a range of 255-a
combination of if .. then .. else and case statements might be more
efficient.)
If we insert this code into Convert, then we might rewrite our loop
to look like this:
{ $1-} program Convert;
begin { main body of Convert }

repeat
Write('Enter input file: '); ReadLn(1nFlleNarne);
Ass:lgn(1nFlle,InFlleName);
Reset(1nFlle); IOCheck
untU DOt 10err;
end. { of program Convert }
Since you are now using IOcheck here (and, presumably,

elsewhere throughout your code), you can now turn off lID error
trapping for the entire program with a single { $I-} directive at the
start ofthe program. Well, you and I have journeyed far. We have
covered many, many aspects of programming using TURBO
Pascal, and with some practice and practical application, you
should be well on your way to solving some real problems of your
own.
I've included a program called GAME1.PAS that brings many of
these concepts together in a practical way, and lets you have fun
besides. You should study it, compile it, 9nd play with it for
awhile. Then, when (and if) you feel up to it, you might want to
tackle the advanced programming concepts presented in Part" I.
But now you have the basis for doing just about anything you
want to do. Your only limit is your imagination, and the sky is the
limit...
18-24
!?5jP *
I~
~'
t
sy-p
- The Sky's the Llmlt-
18-25
PART III
ADVANCED TOPICS
IN TURBO PASCAL
Part 3 Chapter 19 Useful Pascal Routines
19. USEFUL TURBO PASCAL ROUTINES
Welcome to Part III! I am glad you made it this far. Since you are
here, I assume that you are interested in some advanced
programming techniques. Well, there's no shortage of those
here. Happy programming!
Here are some program examples designed for advanced func-
tions under TURBO Pascal. They aren't specific to an operating
system (such as MS-DOS or CP/M); instead, they can be used
anywhere. They're designed as utilities to help you put your
programs together.
All of these programs can befound on yourTurbo Tutor diskette.
The file name of each program is given in the title for each
section. These programs can all be compiled and executed
immediately; no special hardware or software is assumed.
19.1 FUNCTION KEY DETECTION (FUNCKEYS.PAS)
This routine allows you to see what character or character

sequences are produced by any special keys on your computer's
keyboard. This will help you to use those keys for special
functions in your programs. It does assume that any two-
character sequences have ESC (Chr(27)) as the first character.
program GetFunctionKeyData;
{
This program looks at keyboard input to see if a key was hit that
generates a twcrcharacter sequence. On most keyboards, these two
characters consist of chr(27) [ESC] plus an alphanumeric char.
var
Ch : Char;
Previous : Boolean;
Count : Integer;
19-1
Chapter 19 Useful Pascal Routines PartS
begin
far Count:= 1 to 20 do begin
Rea.d(Kbd,Ch); {Reada character, 1fESC (chr(27) then}
if (Ch = chr(27)) aDd. keypressed thaD. begin
{ keystrokB must be either ESC key or one }
Previous := True; { that generates a two-digit code}
Rea.d(Kbd,Ch);
8D4
elM Prev10us := Fa.lse;
ifPrev10us
thaD. Write(/previous ')
elM Write( /s1ngle char ');
Writ&,n(/Ord(Ch)= I ,Ord(Ch))
8D4 end. { of program GetFunct1onKeyDa.ta. }
19.2 BUFFERED INPUT (TVPEAHED.PAS)
This program shows how to guarantee that any characters typed

by the user will be read. I n other words, even if the program is off
doing something else, characters typed will not be lost {up to the
size of the buffer, the size of which depends upon your computer
model and operating system}.
{$O"-,o- }
program Buffered;
{
The $C-d1rective is necessary for type-ahead (buffered)
input; otherwise, characters will be lost. Also, since
Read(Ch) require an end-of-line before processing, you
must use the Read(Kbd,Ch). Type "#,' to end the program.
var
Ch : C~
Indx,Jndx : Integer;
beg:in
reped
far Indx := 1 to 10000 do { delaor loop to show type-ahead }
Jndx := Indx + Indx;
Rea.d(Kbd,Ch); { get next character from buffer}
Write(Ch) { and echo it back out to the screen }
until Ch = I#, { continue until "#,' is entered}
8114. { of program Buffered }
19-2
Part 3 Chapter 19 Useful Pascal Routines
This program will delay between each character. Therefore, it

could take a while to exit from if you typed many characters
ahead.
19.31/0 ERROR CHECKING (IOERROR.PAS)
This program contains a subroutine (IOCheck) along with two

global variables (10 Val and 10Err) that should be a part of any
program which you want to make "bullet-proof". You need to
turn off I/O error trapping (using the $1- compiler option) and
then makes calls to 10Check after most (if not all) I/O operations.
There are two important reasons for this. First, you (and your
program) will then know if any errors have occured and can take
appropriate steps. The global variables 10Val and 10Err can be
used to test for specific problems. Second, if an error has
occured, you need to read 10result (which 10Check does) to
clear it before performing another I/O operation; otherwise, a
second (and different) error can result.
{ $I- }
program TestIOCheck;
{
The routine IOCheck, along with the global declarations
IOFlag and IOErr, should be placed in any program where you
want to handle your own I/O error checking.
}
ccm.st
IOVal : Integer = 0;
IOErr : Boolean = False;
var
InFlle : Text;
Line : striDg[80];
procedure IOCheck;
{
This routine sets IOErr equal to IOresult, then sets
IOFlag accord1.n.gJy. It also prints out a message on
the 24th line of the screen, then waits for the user
to hit any character before proced1ng.
var
Ch : Char;
19-3
Chapter 19 Useful Pascal Routines Part 3
begin
IOVal := IOresult;
IOErr:= (IOVal < > 0);
GotaXY(1,24); ClrEol; { Clear error line in any case}
if IOErr then begin
Write( Chr( 7));
case IOVal of
$01 Write('File does not exist');
$02 Write('File not open for input');
$03 Write('File not open for output');
$04 Write( 'File not open');
$06 Write('Can"t read from this file');
$06 Write('Can"t write to this file');
$10 Write('Error in numeric format');
$20 Write('Operation not allowed on a logical device');
$21 Write('Not allowed in direct mode');
$22 Write('Asslgn to standard files not allowed.');
$90 Write( 'Record length mismatch');
$91 Write('Seek beyond end of file');
$99 Write('Unexpected end of file');
$FO Write('Disk write arror');
$F1 Write('Directory is full');
$F2 Write('File Size oveI'flow');
$FF Write('File disappeared.')
else Write('Unknown I/O error: ',IOVal:3)
end;
Read(Kbd,Ch)
Gli.d
end; { of proc IOCheck }
procedure PutLineNum(LineNum : Integer);

{
This routine tells you which line is being executed,
so that you can see which statement is causing which
error.
}
begin
GotaXY(1,1); ClrEol;
Write( 'Executing line #' ,LineNum)
end; { of proc PutLineNum }
begin
PutLineNum( 1); Assign(InFile,' d:ununY); IOCheck;
PutLineNum(2); Rewrite(InFile); IOCheck;
PutLineNum(3); Read(Infile,Line); IOCheck;
PutLineNum( 4); Close(Infile); IOCheck
end.. { of program TestIOCheck }
19-4
Part 3 Chapter 30 MS-DOS Routines
20. MS-DOS ROUTINES
-DOS Function Cal/s-
Over half of the copies of Turbo Pascal out there are running
under MS-DOS, so this advanced section will start with some
code designed to aid in interfacing with MS-DOS. These routines
are meant to serve as examples for advanced programmers; by
their very nature, they can get you into trouble if you're not sure of
what you're dOing.
All of these programs can be found on your TURBO Tutor disk
(assuming that you have the MS-DOS version). The file name of
each program is given in the title for each section. These
programs can all be compiled and executed immediately; how-
ever, you should look at each program before running it, since it
20-1
Chapter 20 MS-DOS Routines Part 3
may expect certain things, such as a color graphics card or a

software driver.
20.1 RANDOMIZE (RANDOM.PAS)
We'll start with fixing a problem in TURBO Pascal. As you may

have noticed, the procedure Randomize does nothing at all in
TURBO. Here is a procedure for MSand PC-DOS that replaces
the built in Randomize with a working version.
The new Randomize has two Integer parameters. If they are both
0, then the random number seed is set randomly. If either of the
parameters is nonzero, then they are both stored directly into the
32 bit seed.
To set the seed randomly (Randomize(O,O)), the procedure calls
MS-DOS to get the current time. This is a 32 bit value, which is
also stored directly into the seed. On some systems, (Le. the NCR
Decision Mate V), the clock does not tick, so the time never
changes. Randomize checks this, and if the clock hasn't changed
after a Oe/ay(1 00), it asks the user to hit a key. While waiting for
the key, it continuously increments two counters. These are then
stored into the seed.
Please note: This routine is for MS-DOS/PC-DOS TURBO
ONLY!
program. RandomTest;
{
This program tests out the Randomize procedure. It also
calculates a Chi-square value as a test of Random itself:
Chi-square values between 3 and 16 are desirable, with
values close to B.3 being optimum.
var
81,82,Indx,Jndx,Count : Integer;
8um,T,NP : Real;
Tally : array[ 0 ..9] of Integer;
procedure Randomize(I,J: Integer);
20-2
var
BBet : record.
AX,BX,CX,DX,BP,8I,DI,DS,ES,Flags: Integer;
end;
Ch : Char;
begin
if (1=0) aDd (J=O) then begin { Generate a random random number
seed}
RBet.AX:=$2COO; { DOS time of d.aor :function}
MSDos(BBet);
I:=BBet.CXj { Set I and J to the system time }
J:=BBet.DX;
DelacY(l00); {This dalBO' IruW have to be 1ncreasedfor faster systems}
MSDosCBBet);
if (I=BBet.CX) aDd (J=BBet.DX) then begin { Clock isn't ticking}
I :=0;
J:=O;
while KeyPressed do
Read(Kbd,Ch)j { Clear keyboard buffer}
Write( 'Hit any key to set the random number generator: ');
repeat
I := 1+13;
J:=J+17
until Keypressed;
Rea.d(Kbd,Ch)j {Absorb the character}
WriteLn
end
end;
MemW[DSeg:$0129]:=I; {This is the core of the routine: store a 32 bit}
MemW[DSeg:Ol2B]:=J; {seed at locations DSeg:$0129 ...DSeg:$012B}
end; { of procedure Rando:m:ize }
begin{ main be>qy of program RandomTest }

Writeln('Enter count <= 0 to 8D4 program);
repeat
Write( 'Enter count: '); { get # of samples}
ReadLn(Count);
if Count> 0 then begin { do random number test}
Write('Enter seeds (Sl S2): '); { get 2 integers for seed }
ReadLn(Sl,82);
Rando:m:1Ze(Sl,82); { set random number seed }
FillChar(Tal1y,8izeOftTal1y),0); { clear tally arrB(f }
for Indx:= 1 to Count do begin { generate Count numbers}
Jndx := Random(10); { range is 0 ..9 }
Tal1y(Jndx] := Tal1y(Jndx] + 1 { count how many of each }
end;
20-3
Sum:= O.Oj {clear sumforXA 2}

NP := Count/lO.Oj { theoretical number for each}
for Indx := a to 9 do begin { for each possible result do }
Write(Tal1y[IndxJ:5)j {write total for that value}
Sum:= Sum + Sqr(Tal1y[Indx]-NP)/NP { and calculate X A 2 }
end.;
WriteLn;
WriteLn( 'Chi-Square (9 degrees of freedom) = ',8um:8:3)
end
untU Count <= a
end. { of program RandomTest }
20.2 READ DIRECTORY (DIRECTRY.PAS)
TURBO Pascal gives you a lot of control over disk files, including
the ability to erase files as well as to rename them. But nothing
quite beats being able to read the directory off the disk.
Here's a simple program that will read the names of all the files off
a given drive. You can also specify a mask, so that you only get
files matching a given pattern.
program DirListj
{
This is a simple program to list out the directory of the
current (logged) drive.
}
type
Char12arr = array [ 1..12 ] ofCharj
Str:lng20 = str1ng[ 20 ] j
-RegRec =
record
AX, BX, OX, DX, BP, SI, DI, DS, ES, Flags : Integer;
end.;
var
Regs : RegRecj
DTA : array [ 1..43 ] of Bytej
Mask : Char12arrj
NamR : striDg20j
Error, I : Integerj
begin { main body of program DirList }
20-4
F1llChar(DTA,SizeOf(DTA) ,0); { Initialize the DTA buffer}

F1llChar(Mask,SizeOf(Mask) ,0); { Initialize the mask }
F1llChar(NamR,8izeOf(NamR) ,0); { Initialize the flie name }
WriteLn( 'Directory liBt program for MS-Dos.' );

WriteLn;
Regs.AX:= $IAOO; { Function used to set the DTA }
Regs.DS := Seg(DTA); { store the parameter segment in DS }
Regs.DX:= OfsCDTA); {" ""offset in DX}
MSDos(Regs ); { Set DTA location}
Mask:= '???????????'; { Use global search}
Regs.AX:= $4EOO; { Get first directory entry }
Regs.DS := Seg(Mask); { Point to the file Mask}
Regs.DX:= OfsCMask);
Regs.ex := 22; { Store the option}
MSDos(Regs); { Execute MSDos call }
Error := Regs.AX and $FF; { Get Error return}
1:= 1; { initialize 'I' to the first element}
if (Error = 0) then
repeat
NamR[I] := Chr(Mem[Seg(DTA):OfsCDTA)+29+I]);
1:= 1+ 1;
untUnot (NarnR[I-l] in [' ' ..'~']) or (1)20);
NamR[O] := Chr(I-l); { set string length because assigning }
{ by element does not set length}
whlle (Error = 0) do begin
Error:= 0;
Regs.AX := $4FOO; { Function used to get the next }
{ directory entry }
Regs.ex:= 22; { Set the file option}
MSDosC Regs ); { Call MSDos }
Error := Regs.AX and. $FF; { get the Error return}
1:= 1;
repeat
NarnR[I] := Chr(Mem[Seg(DTA):OfsCDTA)+29+I]);
1:= 1+ 1;
untUnot (NarnR[I-l] in [' ' .. '~'] ) or (I> 20);
NamR[O] := Chr(I-l);
if (Error = 0)
then WriteLn(NarnR)
end
end. { of program DirList}
20-5
20.3 READ DIRECTORY II (QDL.PAS)
Here's a more extensive directory-reading program.

»rogram QDL;
»rogram QDL Version 2.00A 09/01/84

QDL uses MSDos to get a listing of an IBM formated diskette.
The function calls used can be found in the DOS Technical Reference
Manual. This program saves the current Data Transfer Area ( DTA) :in
thevariablesDTAseg andDTAofs. The DTA is then reset to the Segment
and Offset of a Buffer variable 'DTA'.
{ $I-,U-,C-}
'tyJIIe { TYPE declarations }

Registers =
record { register pack used :in MSDos call }
AX, BX, OX, DX, BP, SI, DI, DS, ES, Flags: Integer;
end;
CharBOarr = array [ 1 ..80 ] of Char;
Str1ng80 = str1Dg[ 80 ];
var { VARIABLE declarations }
DTA: array [ 1..43 ] of Byte; { Data Transfer Area Buffer}
DTAseg, { DTA Segment before execution}
DTAofs, { DTA Offset" "}
SetDTAseg, { DTA Segment and Offset set after }
SetDTAofs, { start of program }
Error, { Error return}
I,J, { used as counters}
Option: Integer; { used to speci:fY file types}
Regs: registers; { register pack for the DOS call }
Buffer, { generic Buffer}
NarnR : string80; {file name}
Mask: CharBOarr; {file Mask}
20-6
{-------------------------------------------------
SetDTA resets the current DTA to the new address spec1.fied in the
parameters 'SEGMENT' and 'OFFSET'.
procedure SetDTA( Segment, Offset: Integer; var Error: Integer );

begin
Regs.AX := $lADO; { Function used to set the DTA }
Regs.DS := Segment; { store the parameter Segment in DS }
Regs.DX := Offset; {" " " Offset inDX}
MSDos( Regs ); { Set DTA location}
Error := Regs.AX and $FF; { get Error return}
end; { of proc SetDTA }
{--------------------------------------------------
GetCurrentDTA is used to get the current Disk Transfer Area ( DTA )
address. A function code of$2F is stored in the high Byte of the AX register
and a call to the predefined procedure MSDos is made. This can also be
accomplished byustngthe "Intr" procedurewtth the same register record
and a $21 spec1.fication for the interrupt.
procedure GetCurrentDTA( var Segment, Offset: Integer;

var Error: Integer );
begin
Regs.AX := $2FOO; { Function used to get current DTA address}
{ $2FOO is used instead of $2F shl 8 to save
three assembly instructions. An idea for
optimization. }
MSDos( Regs ); { Execute MSDos function request}
Segment := Regs.ES; { Segment ofDTA returned by DOS}
Offset := Regs.BX; { Offset of DTA returned}
Error := Regs.AX and $FF;
end; { of proc GetCurrentDTA
{----------------------------------------------------------------------------------------------------------------
GetDption returns the code used to find the file names on the current
directory ( ie. hidden, standard, or directory ).
procedure GetDption( var Option: Integer );

var
Ch: Char;
begin
Ch:= '?';
Option:= 1;
while ( Ch = '?' ) do begin
20-7
Chapter 20 MS-OOS Routines Part 3
Write( 'File Option to use, [ ? ] for list: ' );

ReadLn( Ch );
WriteLn;
case (Ch) of
'1' : Option:= 1; { -\
'2' : Option:= 7; {\
'3' : Option:= 8; { - These are the options.
'4' : Option := 16; { - Look below for an expla-
'5' : Option := 22; { / of each
'6' : Option:= 31; {- /
'?' : bogin { gives list of possible options
WriteLn( 'File options are: ');
WriteLn;
WriteLn( , [ 1 ] for standard files [ default ].' );
Write( , [ 2 ] for system or hidden files ' );
WriteLn( 'and standard files.' );
WriteLn( , [ 3 ] for volume label.' );
Write( , [ 4 ] for directories and ' );
WriteLn( 'standard files.' );
WriteLn( , [ 5 ] for directories, hidden or ' );
Write( , system files, and standard' );
WriteLn( , files.' );
Write( , [ 6 ] same as 5, but with volume' );
WriteLn( , label included' );
WriteLn;
end;
0100 Option: = 1; { if nothing is typed or an }
Gnd; { incorrect entry is made Option 1 is used}
end
6ll4; { of proc GetDptlon }
{--------------------------------------------------
GetFirst gets the first directory entry of a particular file Mask The Mask is
passed as a parameter 'Mask' and,the Option was previously specified in
the SpectiYQption procedure.
proceduro GetFirst( Mask: Char80arr; var NamR : Str1ng80;

Segment, Offset: Integer; Option: Integer;
var Error: Integer );
var
I : Integer;
begin
Error:= 0;
Regs.AX:= $4EOO; { Get first directory entry }
Regs.DS:= Seg( Mask); { Point to the file Mask}
Regs.DX:= Ofs( Mask);
20-8
Regs.OX := Option; { Store the Option}

MSDos( Regs); { Execute MSDos call }
Error := Regs.AX and $FF; { Get Error return}
1:= 1; { initialize 'I' to the first element}
repeat { Enter the loop that reads in the}
{ first file entry }
NamR[ I ] := ChI'( mem[ Segment: Offset + 29 + I ] );
1:= I + 1;
untll ( not ( NamR[ I - 1 ] in [ , ' ..'-' ] ));
NamR[ 0 ] :=ChI'( I - 1 ); {setstrtnglengthbecauseassigning}
{ by element does not set length}
end; { of proc GetFirst }
{-----------------------------------------------------
GetNextEntry uses the first bytes of the DTA for the file Mask, and returns
the next file entry on cliBk corresponding to the file Mask
procedure GetNextEntry( var N amR : String80; Segment, Offset: Integer;

Option: Integer; var Error: Integer );
var
I : Integer;
begin
Error:= 0;
Regs.AX := $4FOOj { Function used to get the next }
{ directory entry }
Regs.OX := Option; { Set the Ille option}
MSDos( Regs ); { Call MSDos }
Error: = Regs.AX and $FF; { get the Error return}
1:= Ij
repeat
NamR[ I ] := ChI'( mem[ Segment: Offset + 29 + I ] );
1:= 1+ 1;
untU( not ( NamR[ I - 1 ] in [ , ' ..'-' ] ));
NamR[ 0 ]:= ChI'( I - 1 );
end; { of proc GetN extEntry }
main body of program QDL
20-9
begin
for I := 1 to 21 do DTA[ I ] := 0; { Initialize the DTA Buffer}
for I := 1 to 80 do begin { Initialize the Mask and }
Mask[ I ]:= ChI'( 0 ); { file name buffers}
NamR[ I ]:= ChI'( 0 );
end;
NamR[ 0 ]:= ChI'( 0 ); { Set the file name length to 0 }
WriteLn( 'QDL version 2.00A' );
WriteLn;
GetCurrentDT.A( DTAseg, DTAofs,
Error ); { Get the current DTA address}
if ( Error < > 0 ) then begin { Check for errors}
WriteLn( 'Unable to get current DTA' );
WriteLn( 'Program aborting.' ); { and abort. }
Halt; { end program now}
end;
SetDTAseg:= Seg( DTA );
SetDTAofs := Ofs( DTA );
SetDT.A( SetDTAseg, SetDTAofs,
Error ); { Reset DTA addresses}
if ( Error < > 0 ) then begin { Check for errors}
WriteLn( 'Cannot reset DTA' ); { Error message}
WriteLn( 'Program aborting.' );
Halt; { end program }
end;
Error:= 0;
Buffer[ 0 ] := Chr( 0 ); { Set Buffer length to 0 }
GetDption( Option); { Get file Option}
if ( Option < > 8 ) then begin
Write( 'File Mask:' ); { prompt}
ReadLn( Buffer );
WriteLn;
end;
if ( length( Buffer) = 0 ) then { if nothing was entered}
Buffer:= '???????????'; { then use global search}
for I := 1 to length( Buffer) do { Assign Buffer to Mask}
Mask[ I ]:= Buffer[ I ];
GetFirst{ Mask, NamR, SetDTAseg, SetDTAofs, Option, Error );
if ( Error = 0 ) then begin { Get the first directory entry }
if ( Option < > 8 ) then begin { if not volume label }
WriteLn( 'Directory of: " Buffer );{ Write directory message}
WriteLn;
end;
WriteLn( NamR )
end
else if ( Option = 8 ) then
WriteLn( 'Volume label not found' )
20-10
else WriteLn( 'File "', Buffer, "' not found.' )j

while ( Error = 0 ) do begin
GetNextEntry( NarnR, SetDTAseg, SetDTAofs, Option, Error );
if ( Error = 0 ) then WriteLn( N amR );
end;
SetDTA( DTAseg, DTAofs, Error );
end. { end Main }
20.4 DISK STATUS (DISKSTUS.PAS)

This program reads information about free space on a given
drive. Note that the function DefaultDrive can be used to
determine the current logged drive.
program ShowD1skStatus;
.{
5howD1skStatus uses MSDos and the functions therein to get
Drive information on either the current Drive or the Drive
specified on the command line.
}
{ $I-,U-,C-}
ty»e { TYPE declarations }
RegRec=
record { register pack Used in MSDos call }
AX, BX, ex, DX, BP, SI, DI, DS, ES, Flags: Integer;
end;
var
Tracks, { number of available Tracks }
TotalTracks, { number of total Tracks }
Drive, { Drive number}
Bytes, { number of Bytes in one sector}
Sectors : Integer; { number of total Sectors}
Used.,TotalBytes : Real;
Regs : RegRec;
procedure D1skStatus( Drive: Integer;var Tracks, TotalTracks,
Bytes, Sectors: integer);
makes MSDos call to read status of Drive; returns data

in Tracks, TotalTracks, Bytes, and Sectors
20-11
b~in
Regs.AX := $3600; { Get Disk free space}
Regs.DX:= Drive; { Store Drive number}
MSDos( Regs); { Call MSDos to get disk info }
Tracks := Regs.BX; { Get number of Tracks Used}
TotalTracks:= Regs.DX; {" " "of total Tracks }
Bytes := Regs.OX; {" " "of Bytes per sector}
Sectors: = Regs.AX { " " of Sectors per cluster}
end; { of proc DiskStatus }
function DefaultDrive : Integer;

{
makes MSDos call to find out what current default drive is
vnr
Regs: RegRec;
bogin
Regs.AX := $1900; { Get current Drive number}
MSDos( Regs ); { Call MSDos }
DefaultDrive := (Regs.AX and $FF) + 1 { Return value via function}
end; { offunc DefaultDrive}
begin { main body of program ShowDiskStatuB }

Drive := 0; { Initialize Drive}
if (Mem[Cseg:$80J) > 0 {Get command line }
then Drive := Mem[Cseg:$82] and$lF;
if not ( Drive in [ 1 ..6 ]) {If nothing on command line}
thml Drive := 0; { or bad. drive specified,}
{ then logged drive}
DiskStatus( Drive, Tracks, TotalTracks, Bytes, Sectors );
WriteLn;
WriteLn;
Write(' ');
if ( Drive = 0 ) then
Drive:= DefaultDrive;
WriteLn( 'DSCST on Drive ',chr( Drive + $40), ':');
WriteLn;
{ Write disk information}
WriteLn( Tracks:7, 'available tracks.' );
WriteLn( TotalTracks:7, 'total tracks.' );
Used :=(( TotalTracks - Tracks) / TotalTracks) * 100;
WriteLn( Used=7:2, '% used' );
WriteLn( Sectors:7,' sectors per cluster.' );
WriteLn( Bytes:7,' bytes per sector.' );
TotalBytes := (( Sectors * Bytes * 1.0 ) * Tracks );
WriteLn( TotalBytes:7:0, ' total bytes available on disk' );
WriteLn
end. { of program ShowDiskStatuB }
20-12
20.5 DOS VERSION NUMBER (VERSION.PAS)
Here's a much simpler routine, designed to allow you to read the

DOS version number:
program. DosVersion;
function DosVer : Real;

var
Regs : record
AX,BX,CX,DX,BP ,8I,DI,DS)}S,Flags : Integer;
end;
AL,AH : Byte;
begin
Regs.AX := $3000;
MSDos( Regs);
AL:= Regs.AX and$FF;
AH := ( Regs.AX and $FFOO ) ohr 8;
DosVer := AL + AH/l00;
end; { offunc DosVer}
begin { main body of program DosVersion }

WriteLn;
WriteLn;
WrlteLn( DosVer:4:2 );
end. {of program DosVersion}
20.6 DIRECT VIDEO OUTPUT (IBMINT10.PAS)

This routine uses Interrupt 10 to write directly to the screen. This
allows you to do fast, unfiltered screen updates. It is designed to
work only on the I BM PC. It is not generic.
program. IBMpcScreen;
{This program shows the use of direct Video on the IBM-PC ( only! I!)). }
{Interrupt 10 is used for all Video I/O }
20-13
COD8t
Video =$ 10; { Set Video I/O Interrupt}
SetVideo 0; { Set Video mode}
SetCurPosition =$200; { Set cursor position}
ReadCursor = $300; { Read cursor position}
WriteChar =$EOO; { Write character to sceen }
VideoBW80x25A 2; { Mode 80x25 B/W, Alpha }
VideoColorBOx25A 3; { Mode 80x25 Color, Alpha }
type
Result = { Regtster pack}
record
AX,BX,CX,DX,BP,8I,DI,DS,ES,Flags: Integer;
end;
var
Rec :Resul~
Row,Col : Integer;
begin
Rec..AX := SetVideo + VideoBW80x25A; {assumes BW 80x25 displSlY}
Rec.BX := 15;
Intr(Video,Rec); { Set Video Mode}
Rec.AX:= WriteChar + Ord{'A');
Intr(Video,Rec); {Output 'A'}
Rec.AX := ReadCursor;
Intr(Video,Rec); { Read the cursor position}
Row := Rec.DX and. $FFQO ahr 8;
Col := Rec.DX and. $FFj
Write('Row = ';Row,' Column = ',Col); { Show the Row and column}
RecAX := SetCurPosition;
Rec.DX := $DAOA;
Intr(Video,Rec)j { Set the cursor to Row 10 and column 10 }
Rec.AX:= WriteChar + Ord{'#')j
Intr(Video,Rec)j {Output '#'}
Rec.AX := ReadCursor;
Intr(Video,Rec); { Read the cursor position}
Row := Rec.DX and. $FFQO shr 8;
Col := Rec.DX and. $FF;
Write( 'Row = ';Row,' Column = ',Col); { Show the Rowand column }
Rec.AX:= SetCurPosition;
Rec.DX := $1414;
Intr(Video,Rec); {Setthe cursor to Row 20 and column 20 }
end. { of program IBMpcScreen }
20-14
-Interrupts-
20.7 DIRECT MEMORY OUTPUT (MEMSCREN.PAS)
Here's a second way to write directly to the screen: direct

memory addressing. The following program shows how. It's set
up for the Color Graphics display, but can be easily changed to
work with the Monochrome by changing the Segment and Offset
addresses used in the subroutine X. Again, this program is forthe
IBM PC only; it is not generic.
20-15
program ScreenMap;
{program to Write to the screen map}
cODBt
ColoIBeg = $BOOO;
ColorOfs = $8000;
var
I,J : Integer;
C : Char;
RowNumber,ColNumber,Length,Dir
: Integer;
Direction : Char;
procedure X(J,K : Integer);

begin
Mem[ ColoIBeg:ColorOfs + J] : = K { Write to screen memory}
end; { of proc X }
procedure Y(R,C,N,Ch,D : Integer); {r=row C=column n=length ch=char}
var
I,J: Integer;
begin
J :=((R-l)*160) + ((C-l)*2); { compute starting location}
for I := 1 to N do begin
X(J,Ch); { loop n times}
ifD=O
thenJ:= J + 160
elseJ:= J+2;
end
end; { of proc Y }
begin
CIIBcr;
GotoXY(20,10);
Write( 'Do you want to plBiY <YIN> ? ');
Rea,d(kbd,C);
if (C = 'y') or (C = 'Y') then begin
repeat
CIIBcr;
GotoXY(20,10);
Write('Wouldyou like to draw a line <YIN>? ');
Read(kbd,C);
until (C = 'y') or (C = 'Y') or (C = 'n') or (C = 'N');
if (C = 'y') or (C = 'Y') then begin
repeat
CIIBcr;
Y(1,1,80,31,1);
20-16
Y(2,1,24,16,0);
Y(2,80,24,17,0);
Y(26,2,78,30,1 );
GotoXY(20,10);
Write(/The line starts at what row? ');
Read(rownumber);
GotoXY(20,12);
Write(' what colunm? ');
Read( colnumber);
GotoXY(20,14);
Write(' Line length in number? ');
Read(length);
GotoXY(20,16);
Write(' Whatcharactertouse? ');
Read(Kbd,C); Write(C);
repeat
Got£lXY( 10,18); Write(' wh1chd1rect1on<d>ownor<a>cross? ');
Read(Direction)
until (Direction=/a/) or (Direction=/d');
if Direction=, a'
then Dir:=l
elae Dir: =0;
CI:rt3cr;
Y(rownumber,colnumber,length,ord( C),dir);
GotoXY(40,24);
Write('Your line. Try again? ');
Read(Trm,C);
untU(C< > 'Y/) 8D4(C< > 'y);
CI:rt3cr
end
end
elae beg:lD.
GotoXY(23,13);
WriteLn('******** Bye ********/)
end
end. { of program ScreenMap }
20.8 READING THE COMMAND liNE (CMDlINE.PAS)

This next routine can come in very handy. It reads the command
line after the name of your program. This allows you to pass
parameters for your program to use. For example, you could
write
20-17
A>convert stars.txt starS.dat

and pick up the string 'starts. txt starts.dat' to determine what files
to use.
program CommandLine;
This program demonstrates how to get information off the command
line. One thing you must remember- 32 characters are alwBlVs
there for you to use-if you want to use the full 127, the first
statement in your program must parse the command line and retrieve
the information as any subsequent reads or writes w111 shorten the
command line to 32 characters. }
type
CommandString= striDg[127];
var
Buffer : CommandStr1ngj
CL : CommandString absolute cseg:$80;
begin
Buffer := CL; { read the command line immediately }
Gotoxy(20,12);
WriteLn('I' ,Buffer, 'I');
end. { of program CommandLine }
20.9 SERIAL PORT LIBRARY (COMLIB.PAS)

While the standard file Aux lets you talk through the serial port,
you may want something more powerful or sophisticated. Here
are some routines designed to let you select which port to use, set
the baud rate, set the number of stop bits, set the number of data
bits, and select odd, even, or no parity. This is not a generic
program and is intended only for the IBM PC.
{ $U+}
program ComLibTest;
var
Port,Baud,8topBits,DataBits,Par: Integer;
Message: str1Dg[80];
type
~19=~[19];
20-18
{ A set of routines to enable COMI and. COM2 to be accessed from Turbo

Pascal. The following procedures are meant to be called by your programs:
Ass1gnAux(PortNumber in [1,2]) assignsAux to COMI or COM2
AssignUsr(PortNumber in [1,2]) assigns Usr to COMI or COM2
SetSerial(PortNumber in [1,2 J,
BaudRate in [110,160,300,600,1200,2400,4800,9600J,
StopBits in [1,2 J,
DataBits in [7,8J,
Parity in [None,Even,Odd]) sets the baud rate, stop bits, data
bits, and parity of one of the serial ports.
The arr~s InError and OutError ~ be examined to detect errors. The
bits are as follows:
Bit 7 (128) Time out (no device connected)
Bit 3 (8) Framing error
Bit 2 (4) Parity error
Bit 1 (2) Overrun error
Function Ser1al8tatus(PortNumber in [1,2]) returns a more complete
status:
Bit 16 (negative) Time out (no device connected)
Bit 14 (16384) Transmission shift register empty
Bit 13 (8192) TransmiSsion holding register empty
Bit 12 (4096) Break detect
Bit 11 (2048) Framing error
Bit 10 (1024) Parity error
Bit 9 (612) Overrun error
Bit 8 (266) Datarea,qy
Bit 7 (128) Received line signal detect
Bit 6 (64) Ring indicator
Bit 6 (32) Data set ready
Bit 4 (16) Clear to send
Bit 3 (8) Delta receive line signal detect
Bit 2 (4) Tra1l1ng edge ring detector
Bit 1 (2) Delta data set rea,qy
Bit ° (1) Delta clear to send
Identifiers starting with "__" are not meant to be used by the user
program.
}
type
_ -Registenget=Record case Integer of
1: (AX,BX,CX,DX,BP,DI,8E,DS,ES,Flags: Integer)j
2: (AL,AH,BL,BH,CL,CH,DL,DH: Byte)j
end;
- .YaritfiYpe=(None,Even,Odd)j
20-19
var
_ ...Regs: _ ...RegisterSet;
lnError,OutError: array [1 ..2] of Byte;
procedure _ Jnt14(PortNumber,Command,Parameter: Integer);
{ do a BIOS COM driver interrupt}
begin
with _ ...Regs 40
begin
DX:=PortNumber-1 ;
AH:=Command;
AL:=Parameter;
Flags: =0;
Intr($141-...Regs);
end;
end;
procedure SetBerial(PortNumber,BaudRate,BtopBits,DataBits: Integer;

Parity: _ -ParityI'ype);
{ Set serial parameters on a COM port }
var
Parameter: Integer;
begin
case BaudRate of
110 : BaudRate:=O;
160 : BaudRate:=l;
300 : BaudRate:=2;
600 : BaudRate:=3;
1200: BaudRate:=4;
2400: BaudRate:=6;
4800: BaudRate:=6;
else BaudRate:=7; { Default to 9600 baud}
end;
if StopBits=2 then StopBits:= 1
else StopBits:=O; { Default to 1 stop bit}
if DataBits=7 then DataBits:=2
else DataBits:=3; {Default to 8 data bits}
Parameter:=(BaudRate sh16)+(StopBits sh12) + DataBits;
case Parity of
Odd: Parameter:=Parameter+8;
Even: Parameter:=Parameter+24;
else; { Default to no parity }
end;
_ Jnt14(PortNumber,O,Parameter);
end;
20-20
Function SerialStatuB (PortNumber: Integer): Integer;

{ Return the status of a COM port }
begin
_....Intl4 (PortNumber,3,O);
SerialStatus := _ -.Regs.AX;
end;
procedurL _OutPortl( C: Byte);
{ Called by Write to Aux or Usr when assigned to COMl }
begin
while (SerialStatus( l) and $30)=0 do ;
_....Intl4(l,l,C);
OutError[l]:=OutError[l] Or (_-.Regs.AH and $BE);
end;
procedurL _OutPort2( C: Byte);

{ Called by Write to Aux or Usr when assigned to COM2 }
begin
while (SerialStatuS(2) and $30)=0 do ;
__Intl4(2,l,C);
OutError[2]:=OutError[2] or (_ -.Regs.AH and $BE);
end;
Function _ -InPortl: Char;

{ Called by Read from Aux or Usr when assigned to COMl }
begin
__ Intl4( 1 ,2,0);
__InPortl: =Chr( _ -.Regs.AL);
InError[l]:=InError[l] Or (_-.Regs.AH and$8E);
end;
J'aD.ction _ -InPort2: Chari
{ Called by Read fromAux or Usr when assigned to COM2 }
begin
__Intl4(2,2,0);
__InPort2:=Chr(_ -.Regs.AL);
InError[2]:=InError[2] or (_ -.Regs.AH and $BE)i
end;
20-21
procedure_-ABsignPort(PortNumber: Integer; var InPtr,OutPtr: Integer);

{ Assign either Aux or Usr to either COMI or COM2 }
begin
if PortNumber=2 then
begin
OutPtr:=Ofs(__OutPort2);
InPtr: =Ofs( _ JnPort2);
end
else { Default to port I }
begin
OutPtr:=Ofs( __OutPortI);
InPtr: =Ofs( _ JnPortI);
end;
InError[PortNumber]:=O;
OutError[PortNumber]:=O;
end;
procedure Ass1gn.Aux(PortNumber: Integer);

{AssignAux to either COMI or COM2 }
begin
_ -ABsignPort(PortNumber~uxInPtr~uxOutPtr);
end;
procedure AssignUsr(PortNumber: Integer);

{ Assign Usr to either COMI or COM2 }
begin
_ -ABsignPort(Port1~umber,UsrlnPtr,UsrOutPtr);
end;
Function Binary(V: Integer): StringI9;
var
I: Integer; B: array [0 ..3] ofB'triDg[4];
20-22
begin
for 1:=0 to 16 do
if (V and.(18h1(16-1)))<>O thenB[1 div4][(I mod4)+I]:='I'
else B[1 div4][(1 mod4)+I] := '0';
for 1:= 0 to 3 doB[1][O]:= Chr(4);
B1na.ry:=B[O]+' '+B[I]+' '+B[2]+' '+B[3];
end;
begin
Write( 'Enter port number: ');
ReadLn(Port);
AsslgnUsr(Port);
Write('Enter baud rate: ');
ReadLn(Baud);
Write('Enter stop bitB: ');
ReadLn(StopBitB);
Write( 'Enter data bItB: ');
ReadLn(DataBitB );
Write('Enter parity (O=none, 1 =even, 2=odd): ');
ReadLn(Par);
Write('Enter message to print: ');
ReadLn(Message);
SetSer1a.l( l,Baud,StopBItB,DataBitB,ParltyType(Par));
WrlteLn(Usr,Message);
WrlteLn('OutError[' ,Port,']: ' ,B1nary(OutError[Port]));
WriteLn( 'Ser1aJBtatus(' 'port,'): ' ,B1nary(Ser1aJBtatus(Port)));
end. { of program ComLibTest }
20.10 MICROSOFT MOUSE INTERFACE

(TBOMOUSE.PAS)
The Microsoft Mouse runs off of a hardware/software combina-

tion: a card which plugs into a slot, and a software driver
(MOUSE.COM) which (when executed) installs itself in RAM.
The following program shows how to read the Mouse from Turbo
Pascal. This is not a generic program; it works only on the IBM
PC.
program Mousa<3ketch;
{
This program shows how to read the Microsoft Mouse.
***WARNING***
Be sure that you have loaded the mouse driver
(by runn:1Ilg MOUSE.COM) before executing this program.
20-23
type
Regpack=
record
AX,BX,CX,DX,BP,8I,DI,DS,ES,Flags : Integer;
end;
var
OldX,OldY;X,Y : Integer;
M1,M2,M3,M4 : Integer;
Regpak : Regpack;
procedure MOUBe(var M1,M2,M3,M4 : Integer)j

var
Regs : Regpack;
begin
with Regs do begin
AX := M1j { Set up ax,bx,cx,dx for interrupt}
BX:= M2j
CX:= M3j
DX:= M4
end;
Intr(51,Regs)j {Trip interrupt 51 }
with Regs do begin
M1:= AXj
M2:= BXj
M3:= CXj
M4:= DX
end.
end; { of proc Mouse}
20-24
begin { main boqy of program Mous~ketch }

Ml := 0;
M2 := 0;
M3 := 0;
M4 := 0;
HiRes; { Choose graphics mode and color}
HiResColorCYellow);
Ml := 0; {Initialize mouse driver}
Mouse(Ml,M2,M3,M4);
Ml := 1; {Turn on Mouse cursor}
Mouse(Ml,M2,M3,M4);
Ml := 3;
OldX:= 0;
OldY:= 0;
while not KeyPressed do begin { Exit mouse when any key pressed}
Mouse(Ml,M2,M3,M4);
ifM2<>0
then Draw(OldX,OldY,M3,M4,1); {Draw ifbutton pushed}
OldX:= M3;
OldY:= M4
eDd.
end. { of program Mous~ketch }
20.11 FILLCHAR DEMO (FILLCHAR.PAS)

The FiliChar routine is a handy means of intializing arrays and
other data structures. This program gives an example of how it
works and shows how it fills memory with a given value. It also
shows how to examine (dump) a portion of memoryfrom within a
program.
program DemoF1llChar;
{
This program demonstrates how the built-in procedure F1llChar works
var
Segl: Integer;
Ofsl: Integer;
Count: Integer;
OutWord: Integer;
Num: Integer;
Varl: Integer;
Value: Char;
20-25
Chapter 20 MS-DOS Routines PartS
begin
ClIf3crj
GotoXY(20,5 );
Write('enterValue for starting address')j
ReadLn(Varl )j
Segl := Seg(Varl);
Ofsl := Ofs(Varl)j
WriteLn;
WriteLn( 'This variable is at segment: ' ,segl " with an offset of ',Ofsl);
WriteLn;
WriU('NowputinaValue(si.nglechar)thatyouw1Bhmemoryloadedwith: ');
ReadLn(Value);
Write('Put in how manywordByou want filled: ');
ReadLn(Num);
FillChar(Varl,Num,Value);
ClIf3cr;
WriteLn('Now we print our memory starting with ' ,segl ,':' ,Ofsl);
for Count := 1 to Num do begin
OutWord:= Mem[Segl:Ofsl];
WriteLn(Segl,':',Ofsl,' has value ',OutWord);
Ofsl := Ofsl + 1;
end
end. { of program DemoFillChar }
20-26
Part 3 Chapter 21 CP/M Routines
21. CP/M R'OUTINES
Here are some program examples designed for advanced func-

tions under CP/M. These routines are meant to serve as examples
for advanced programmers; by their very nature, they can get you
into trouble if you're not sure of what you're doing.
All of these programs can be found on your Turbo Tutor disk
(assuming that you have the CP/M version). The file name of
each program is given in the title for each section. These
programs can all be compiled and executed immediately; no
special hardware or software is assumed.
21.1 READ DIRECTORY (CPMDIR.PAS)
This program allows you to read the directory of the current

logged drive. It makes the appropriate BOos calls, pulling the file
names out of the memory location where CP1M leaves them.
program CPM80Dir;
{ This program will give a directory of the logged drive. }
CODBt
SearclLFirst : Integer = $11;
SearclLNext : Integer = $12;
Set...DMA : Integer = $1A;
var
Error, Loop, Start : Integer;
FOB : array[ 0 ..26] of Byte absolute $0060;
DMA : array[ 0 ..266] of Byte;
21-1
Chapter 21 CP/M Routines Part 3
begin
Error := BDos(Set..DMA.,Addr(DMA))j
FCB[O]:= OJ
for Loop := 1 to 11 do
FCB[Loop] := ord('?')j
Error:= BDos(Search..First~ddr(FCB))j
if Error < > 255 then begin
Start : = Error * 32j
for Loop:= Start to start+8 do
Write(Cha.r(Mem[Addr(DMA)+Loop]) )j
Write(' ')j
for Loop:= Start+9 to Start+11 do
Write(Cha.r(Mem[Addr(DMA)+Loop]))j
WriteLn
end;
repeat
Error:= BDos(searclLNext);
Start := Error * 32;
if Error < > 255 then bagin
for Loop:= Start to start+8 do
Write(Cha.r(Mem[Addr(DMA)+Loop]));
Write(' ');
for Loop:= Start+9 to Start+11 do
Write(Cha.r(Mem[Addr(DMA)+Loop]));
WriteLn
end
until Error=255
end. { of program CPM80Dir }
21.2 SYSTEM STATUS (CPMSTAT.DIR)
This program does a more exhaustive displayof system status. It

shows information about each disk drive, as well as the version of
CP/M currently being used. This program is intended for use on
CP/M versions 2.0 or higher.
program. CPMstatus;
{
Reads and d1sp1aors CP1M status information
}
cODBt
CPMversion 12;
CurDiBk 25;
AllocVector 27;
DiBkParam 31;
GetUser 32;
21-2
Part 3 Chapter 21 CP/M Routines
type
Word = Intege~
HexStr = string[ 4];
DPBREC=
record
SPr : Integer; {SEareRS PER TRACK}
BSH : Byte; {DATAALLOCATIONBIDCKSHIFrFAaroR}
BLM : Byte;
EXM : Byte;
(* DSM : Integer; {TOTAL STORAGE CAPACITY} *)
DSMlo : Byte;
DSMhi : Byte;
DRM : Integer; {NO ofDIREarORY ENTRIES}
.AJ.JJ,AL1 : Byte;
OKS : Intege~
OFF : Integer
end; { DPBREC }
var
DPB : A DPBREC;
RecsPrBlock : Integer;
RecsPrDrive : Real;
TrksPrDrive : Real;
BIOSaddr : Integer absolute 1;

BDOSaddr : Integer absolute 6;
TPA : Real;
Version : Intege~
Result : Integer;
:function Hex(Number: Integer;Bytes: Integer): HexStr;

ccmst
T : array[O .. 15] of Char = '0123456789ABCDEF';
var
D : Integer;
H : HexStr;
begin H[O]:=Chr(Bytes+Bytes);
for D:=Bytes+Bytes dawnto 1 do begin
H[D]:=T[Number and 15];
Number:=Number 8hr 4
end;
Hex:=H
end; { ofproc Hex}
21-3
Chapter 21 CP/M Routines PartS
begin { main body of program CPMStatus }

ClIScr;
Writeln(
, -Logged- --RecordB-- -Tracks- -CapaciW- -TPA-')
;wr1teln(
, Drive User Block Track Drive Bys. Drive Directory Drive Bytes K');
{ x: xxx xxxxx: xxxxx xxxxx xxx xxxxx xxxx/xxxx xxxxxK xxxxx xx.x}
Write(' ',Chr(Bdos(CurD1Bk) + Ord('A')), ':

Bdos(GetUser,$FFFF):3,' ');
DPB:=Ftr(BdosHJ..(D1BkParam));
with DP&" 40 begin
RecsPrBlock:=BLM +1;
Write(RecsPrBlock:6 );
Write(SPr:6);
RecsPrDrive:=(DSMhi*266.0 + DSMlo + 1.0)*RecsPrBlock;
Write(RecsPrDrive:6:0,' ');
Write(OFF:3);
TrkBPrDrive:=RecsPrDrive/SPr + OFF;
if TrkBPrDrive < > Trunc(TrkBPrDrive)
then TrkBPrDrive:=TrkBPrDrive+1;
Write(Trunc(TrkBPrDrive):7,' ');
Write(DRM +1:4,'1' ,CKS*4:4,Trunc(RecsPrDriveIB):6,'K');
cmc1;
TPA:=2.0*(BDOSacldrshr 1) -$100;
WriteLn(TPA:B:O,' , ,TPA/I024:6: 1);
WriteLn;
Vlriteln('- Operating System -');
Writeln(' Version BDOS BIOS');
{ xxxx x.x xxxx xxxx}
Result:=BdosHJ..(CPMverslon);
Version:=Hi(Result);
itVersion = 0
then Write('CP/M ')
else if Version = 1
then Write('MP/M ')
else Write('???? ');
Verslon:=Lo(Result);
ifVers10n = 0
then Wrlte('I.x')
elae Write(Version div $10, '.', Version mod $10);
WriteLn(Hex(BOOSaddr, 2):6, Hex(BIOSacldr-3, 2):6);
end. { of program CPMStatus }
21-4
PartS Chapter 22 Assembly Language Programming
22. ASSEMBLY LANGUAGE

PROGRAMMING
-Assembler-
There are usually three reasons for calling assembly language
from a high-level language. The first is to execute some system
operation; the second, to perform some sort of data manipulation;
the third, to carry out some set of instructions more quickly. In al!
22-1
Chapter 22 Assembly Language Programming Part 3
three cases, the action is performed in assembly language

because of limitations in the high-level language. Ironically, these
problems don't crop up all that often in TURBO Pascal. First,
TURBO provides broad access to system functions and re-
sources. You can make calls directly to the operating system (via
BOos, BIOS, MSDos, Intr), you can directly access any location
in memory, and you can reference all the I/O ports. Second,
Turbo Pascal provides low-level operations (such as shr and shl)
and also extends logical operations (and, or, etc.) to integer
values. Retyping capabilities bypass the need for assembly
language data conversion functions. Finally, TURBO Pascal
produces machine code which is often fast enough to satisfy
your needs.
22.1 IN-LINE CODE (INLINE.PAS)
The above statement is often true, but not always ... so, yes, there
may be time when you want to write assembly language routines.
In that case, you have two routes to go. The first (and simplest) is
to use in-line machine code. Turbo Pascal allows you to insert
(anywhere!) machine code, which will then be executed everytime
the inline statement is encountered.
The inline statement takes the form:
:ln11ne( vall/vall/.. ./valn);
where each value (vaI1, etc) is a constant (either literal or named,

and of type integer), a variable identifier, or a location counter.
More information can be found in the TURBO Pascal Reference
Manual, in the appropriate appendices foryouroperating system.
Here's asample program, written to run on an 8086-based system
(under either MS-DOS or CP/M-86). The routine does a simple
divide-by-two of the value passed to it; note how the parameter's
name (a variable identifier) can be put right into the inline
statement.
22-2
Part 3 Chapter 22 Assembly Language Programming
program InLinEBample;
{
The following program example divides even integers by two.
For odd integers this program returns -32768 + the value
divided by two (integer division)
var
Value : Integer;
procedure VInLine(var Value:Integer);
{
a simple use of inline code. Note that some constants
have been defined and used, while other values are left
as literal hexadecimal constants. This is done just for
illustration.
}
ccmst
CLC = $F8;
IN_CDI = $47;
begin
iDline
($C4/$BE/VALUE/ {LEB DI;VALUE[BP] }
CLC/ { CLC }
$26/$DO/$lD/ {RCR EB:BYTE PrR [DI] }
IN_CDI! { INC DI }
$26/$DO/$lD); {RCR EB:BYTE PrR [DI] }
end; { of proc VInLine }
begin { main body of program InLinEBample }

CIIScr;
repeat
Write('Enter a number, <0> to quit: ');
ReadLn(Value);
VInLine(Value);
WriteLn('Return Value is; ',Value);
untU Value = OJ
CIIScr;
end. { of program InLinEBample }
22.2 ASSEMBLY LANGUAGE ROUTINES

(PASSFUNC.PAS,PASS.ASM)
TURBO Pascal allows you to call assembly language procedures

and functions that you have assembled separately. These are
22-3
known as external subprograms. If you are running under CP/M-

80, you have to declare these routines to be at specific locations
in memory:
procedure LowToUp(var Str : BigStr); extemal $2EOO;
When your program starts executing, it must go out and itself
load the routine at the appropriate memory location, with code
looking something like this:
procedure LoadRoutine;
var
CodeFlle : File;
Buffer : BrrQ'[0 ..1] otByte absolute $2EOO;
Index,Rae : Integer;
begin
Ass1gn(CodeFlle,'LOWTOUP.COM');
Reset( CodeFile);
Index := 0; Ree := 0;
{ $R-}
whUe DOt EOF(CodeFile) 40 begin
BloekRead( CodeFile,Buffer[Index] ,Rae);
Rae := Ree + 1;
Index := Index + 128
end;
{ $R+}
Close( CodeFile)
end; { of proe LoadRoutine }
If instead you are using MS-DOS, then you must write a true
assembly language routine which is assembled separat,ely and
then linked in at compile time. The advantage is that you can write
directly in assembly language (rather than "hand-assembling"
your routines into machine code for the inline statement). The
disadvantages are that you have to own an assembler (and one
that's compatible with TURBO Pascal), and that you have to
make sure that your assembly routines and your Pascal code can
correctly communicate with each other. Again, the particulars
are specific to each operating system, so you are again referred
to the appropriate appendices in the TURBO Pascal Reference
Manual.
22-4
For those you you with MS-DOS systems, here is a sample

assembly !anguage routine:
; * * WARNING * WARNING * WARNING * WARNING * WARNING *

; Please do not try to use external functions
; unless you are familiar with assemb1y language.
,
; IMPORTANT: Externals must be written in 8Bsemb1y language.
,
j The follaw1ng example addes two integer numbers.
code segment
assume cs:code
pass proc near
push bp j SAVE ENVIRONMENT
mov bp,sp
inov ax,[bp+4] j GET PARAMETER 1
add ax,[bp+6] ; GET PARAMETER 2
j GIVES THE RESULT
mov sp,bp j RESTORE ENVIRONMENT
pop bp
ret 4
pass endp
code ends
end
j Now exit to PC-ooS and t:Q7pe:
>ASMPASS
>LINKPASS
>EXE2BIN PASS.EXE PABS.COM
,
; Ignore minor errors from ABM and LINK.
Having carried out the steps above, you can now compile the
following TURBO Pascal program (making surethat PASS.COM
is on the default (logged) drive):
»rogram PassFunc;
{
ThiS routine expects the assemb1y language routine 'Pass'
to reside in the file 'PASS.COM'
var
Varl, Var2, Var3: Integer;
22-5
faD.ctionPass(VarX, VarY: Integer): Integer; e.xtemal 'PASS.COM';

begin { main boqy of program PassFunc }
repeat
ReadLn(Varl );
ifVarl< >0 then begin
ReadLn(Var2);
Var3 := Pass(Varl,Var2);
WrlteLn(Varl,' + 'Nar2,' = ',Var3);
WriteLn
end
until Varl = 0
end.. { of program PassFunc }
When this program compiles, TURBO Pascal will look forthe file
PASS.COM and link the machine code into the executable code
being produced.
22.3 CONCLUSION
Well, this is the end of my little book. I have enjoyed telling you
what I know about TURBO Pascal. I hope you've enjoyed
learning about it. All of the tools are now in your hands. It is up to
you to use those tools like any skilled craftsman would. Practice
what you have learned until you master it, then move on to
something else.
Since you have read the entire manual, I would like to reward you
with a few bits of wit I thought you might enjoy:
1. If it were not for the last minute, nothing would ever get done.
2. The efficency of a programming team is inversely propor-
tional to the number of members in that team, whenever that
number is greater than three!
3. Even the simplest things are more complicated than you
thought.
4. The first 90 percent of a job takes 90 percent of the available
time; the last 10 percent takes the other 90 percent.
5. Always plan for the project to take twice as long as you
expected, after first doubling your original estimates.
22-6
Thank you for your attention to the matters of programming in

TURBO Pascal and for buying this book. If you are programming
for your own entertainment, your newly-learned programming
skills are probably sufficient for anything you will want to do. If
you intend to write serious Pascal programs, there are a couple of
other books I strongly recommend. These are:
-Wirth, Niklaus: Algorithms + Data Structures = Programs.
Prentice Hall (1976).
-Wirth, Niklaus: Pascal. Prentice Hall.
-Knuth, Donald E.: The Art of Computer Programming. (Vols.
1,2, and 3). Addison Wesley (1973).
-Horowitz, E. et al: Fundamentals of Data Structures. Pitman
(1976).
22-7
NOTES
INDEX
A comment delimiters, 8-11

About the author, v commenting out sections of code
8-9 '
advanced program structure, 8-12
array, 13-1 comments, 7-4,8-8
data type, 13-2 compiled code, 5-4
definition, 13-1 compiler
index range, 13-2 options, 6-10
indexing, 13-4 compiling a program, 6-8
initialization, 13-5 compound statement, 10-2
order of. elements, 13-6 conditional execution, 10-4
packed, 13-4 constant, 3-5, 7-9
selecting an element, 13-3 typed,8-14
space considerations, 13-5 hex, 9-7
{$R+} compiler option, 13-4 control characters, 7-6
ASCII character entering, 9-16
table of, 7-13 control structures, 10-1
ASCII characters, 7-6 control structures program
assembly language programming example, 10-15
2~1 ' converting data, 12-11
assembly language routines, 22-3 CPMDIR.PAS routine, 21-1
CPMSTAT.DIR,21-2
B cumulative round-off error, 9-11
block statements, 8-17, 10-2 o

nesting, 8-18
bold data, 4-3
in program listings, 3-3 conversion, 12-11
Boolean data, 9-17 character, 4-3
boolean expressions, 10-3 label,4-3
boolean truth table, 10-15 numeric, 4-3
buffered input routine, 19-2 predefined data types, 8-3
byte manipulation, 9-9 deallocation, 17-9
byte types, 9-9 declaration section, 8-13
const, 8-13
C label,8-13
type, 8-14
case statement, 10-11 var, 8-14
chaining files, 6-9 declaration statement, 8-2
character data, 4-2 declared scalar types, 12-1
character types, 9-15 declaring subranges, 12-13
characters, 7-6 declaring variables, 9-1
non-printing, 7-6 defining a "none of the above"
printing, 7-7 var, 12-6
special,7-7 defining variables, 7-3
CMDUNE.PAS routir,e, 20-17 developing a program, 5-1
COMUB.PAS routine, 20-18
23-1
device 1/0, 18-16 forward declarations, '11-12
direct memory output routine, free union, 15-12
20-15 FUNCKEYS.PAS routine, 19-1
direct video output routine, 20-13 function key detection routine, 19-1
directory display, 6-14 functions, 9-8, 11-1, 11-8
DIRECTRY.PAS routine, 20-3
discriminated union, 15-12 G
disk status routine, 20-11 getting started, 6-1
DISKSTUS.PAS routine, 20-11
div, 9-3 H
DOS version number routine, 20-13
DST heap, 17-4
ordinal value, 12-4 size of, 17-4
reading values, 12-10 use of, 17-5
DSTs, 12-1 hex constants, 9-7
dynamic allocation, 17-1 History of Pascal, 2-1
E
editor, 6-6 1/0 error checking routine, 19-3
EOLN function, 18-9 1/0 error handling, 18-20
equations 1/0 special filenames, 18-18
using parentheses, 9-5 1/0 {$I-} compiler option, 18-15
error handling, 6-9 IBMINT10.PAS routine, 20-13
1/0,18-20 identifiers, 7-8, 9-2
error if...then ... else, 10-11
run-time, 6-12 in-line code routine, 22-2
type mismatch, 9-12 information, 4-2
value range error, 9-6 INLlNE.PAS routine, 22-2
executing a program, 6-16 input/output, 7-2
expression integer expressions, 9-4
boolean, 10-3 integers, 9-3
integer, 9-4 real,9-10
expressions, 7-11 unsigned,9-7
predefined functions, 9-7
F 10ERROR.PAS routine, 19-3
iteration, 10-6
file, 18-1
appending records to, 18-6 K
chaining, 6-12
definition, 18-1 keyboard
directory of, 6-15 real-time input, 18-19
random access, 18-5
saving, 6-14 L
text, 18-7 languages
untyped, 18-14 structured, 2-5
filenames, 18-13 linked list, 17-6
fillchar demo routine, 20-25 use of, 17-8
FILLCHAR.PAS routine, 20-25 local variable, 11-5
for... do loop, 10-7 logical operators, 9-3
formatted output, 18-10 looping, 10-6
23-2
M precedence of operations, 9-5
predefined data types, 8-3, 9-1
main body of a program, 7-3 predefined functions
main menu, 6-3 for integers, 9-7
memory management, 17-9 table of, 8-19
memory space, 6-16 predefined procedure, 8-6
MEMSCREN.PAS routine, 20-115
table of, 8-18
Microsoft Mouse interface routine,
procedure
20-23 definition, 1-4
mixing strings, characters, predefined, 8-6
and arrays, 14-16 procedures, 9-8, 11-1
mod,9-3 program statement, 8-2
program structure
N advanced, 8-11
Niklaus Wirth, 2-5 program
nil,7-10 body, 8-4
nonsequential program execution, commenting, 7-4, 8-8
10-1 conditional execution, 10-4
normalization, 9-14 declaration section, 8-1
not operator, 9-8 header, 3-5
numeric conversions, 14-13 main body, 7-3
numeric data, 4-2 parts of, 3-5
running, 6-10
o simple, 3-3
object code, 5-4 statements, 3-6
operator precedence, 9-5, 9-11 structure, 8-1
operator programming
unary, 9-8 basic concepts, 4-1
operators, 9-8 considering data to use, 5-3
order of elements in an array, 13-6 habits, 3-2 .
order of evaluation, 9-5 identifying the problem, 5-2
ordinal values, 12-4 pseudocode, 5-4
other books to read, 22-7
Q
p QDL.PAS routine, 20-6
packed array, 13-5
parameter lists, 11-8 R
parameters, 11-6 random accesss of files
parentheses definition, 18-6
use in equations, 9-5 RANDOM.PAS routine, 20-2
Pascal randomize rOL!tine, 20-2
converting BASIC to, 7-1 read command line routine, 20-17
history, 2-1 read directory routine, 20-4, 21-1
TURBO vs. Standard, 8-12 reading DST values, 12-10
PASS.ASM routine, 22-3 real integers, 9-10
PASSFUNC.PAS routine, 22-2 real number
pointer, 17-2, 17-3 exponent, 9-13
pointing, 17-4 precision, 9-13
use of, 17-3 round-off error, 9-13
portable languages, 2-3 real-time keyboard input, 18-19
23-3
record, 15-1 subranges, 12-12
field,15-2 swapping values, 11-7
space considerations, 15-10 symbols, 7-9
variant, 15-8 system status routine, 21-2
recursive subprogram, 11-11
T
remainder, 9-3
repeaL.untilloop, 10-10 TBOMOUSE.PAS routine, 20-23
reserved words, 3-3, 7-8 terminal installation program, 6-1
as symbols, 7-9 text files, 18-7
run-time errors, 6-12, 9-13 tinst, 6-1
running a program, 6-10 truth table, 10-15
type mismatch error message, 9-12
S TVPEAHED.PAS routine, 19-2
typed constants, 8-14
saving files, 6-14
scope of identifiers, 11-3 U
semicolon, 3-6 unary operator, 9-8
exceptions, 3-6 union
use of, 7-3, 8-6 discriminated, 15-12
serial port library routine, 20-18 free, 15-12
set, 16-1 Unknown Identifier error message,
comparison, 16-3 9-2
constant, 16-1 unsigned values, 9-7
defining, 16-1 untyped files, 18-14
difference, 16-5 using parentheses in equations, 9-5
intersection, 16-5
operations, 16-4 V
size, 16-2 value range error message, 9-6
union, 16-5 variable, 3-5, 4-4, 7-10, 9-2,13-1
use of, 16-2 Boolean, 9-17
simple word processor, 14-5 byte, 9-9
source code, 5-4 char, 9-15
space creating, 9-1
in memory, 6-16 declaring, 9-1
statement, 3-6, 4-5, 10-1 defining, 7-3
compound, 10-2 entering control characters, 9-16
string, 14-1 forward declaration, 11-12
and {$V-} option, 14-16 integer type, 9-3
,as parameter, 14-14 local, 11-5
comparisons, 14-5 of type string, 14-2
definition, 14-2 previously declared, 11-4
function, 14-7 variant records, 15-8
numeric conversion, 14-13 VERSION.PAS routine, 20-13
procedure, 14-7
W
procedure & function table, 14-18
structured language, 2-5 while ... do loop, 10-9
subprogram, 8-15, 11-1 Who is Frank Borland?, 4
location, 11-2 Wirth; Niklaus, 2-5
recursive, 11-11 WITH statement, 15-4
subrange nesting, 15-5
declaration, 12-13 writing a simple program, 3-3
writing to disk, 18-5
23-4
CATALOG
or ALL
BORLAND
PRODUCTS
.>.> BORlAnD
INTERNATIONAL
NOTES
SUPERKEY TURNS 1,000 KEYSTROKES INTO 1
Bllllrll'
AND KEEPS YOUR 'CONFIDENTIAL FILES' ... CONFIDENTIAL.
Borland International, the folks who brought you SIDEKICKn•., now offer you a keyboard enhancer that
can truly make your life easier and your files safer.
When you've got a lot to do, SUPERKEy™ lets you dramatically cut the number of times you touch
the keyboard. Even your most sophisticated programs becomes one-stroke easy.
SUPERKEY also performs as a high-level security encryptor/decryptor to protect your sensitive files.
Just look at what SUPERKEY'S got!
• HIGH-LEVEL RESIDENT ENCRYPTION: SUPERKEY includes a two-mode encryption system that

uses your secret password to encrypt and decrypt files -even while running other programs.
Mode 1: DIRECT-OVERWRITE ENCRYPTION: Just give SUPERKEY your password and the files of
your choice will be encrypted or decrypted - without changing file size or creating backup
files for even greater document security. You can even encrypt public bulletin board messages.
No more curious reading of your files unless you provide the secret password.
Mode 2: EXE OR COM FILE ENCRYPTION: Using your personal code, SUPERKEY can encrypt your
files for transmission (over a phone line, for instance). On the other end, only your secret
code can reconstruct the file.
• PRIVACY MODE: SUPERKEY lets you turn off your screen and lock your keyboard against intruders.
• SCREEN BURN-IN PROTECTION: SUPERKEY will switch your monitor off if there's no input or screen
110. As soon as a key is touched, the screen image returns.
• RAM RESIDENCE AT ALL TIMES: Create, edit, save and even recall new or existing macro files
anytime, even while running another program.
• PULL-DOWN MACRO EDITING: A single keystroke opens a full-screen WordStarTr"-command edit
window. You can perform most word processing functions including file naming. SUPERKEY even
has our 'clip & tape' function for moving blocks of data from the screen for later retrieval.
• COMMAND KEY REDEFINITION: With a few easy strokes, you can redefine key commands 'on the fly'.
• ENTRY AND FORMAT CONTROL IN DATA FIELDS: SUPERKEY tests for digits or letters, converts
lower case to upper case and visa-versa, and justifies files right, left, or center.
• TOTAL PROKEY COMPATIBILITY: SUPERKEY uses every ProKey ™ macro file without any con
version! No time is wasted and you're that far ahead.
• DISPLAY-ONLY MACRO WINDOW: Create your own customized on-line help system for other users.
An ideal feature to set-up interactive tutorials on your computer.
• CONTEXT-SENSITIVE ON-LINE HELP SYSTEM: You can request just the help you want-no more
wading through the whole operating manual.
SUPERKEY is now available for an unbelievable $69.95 (not copy-protected).

Minimum System Configuration: SUPERKEY is compatible with your IBM PC. XT. AT. jr. and 100% compatible microcomputers. You
computer must have at least 128K RAM. one disk drive and PC· DOS 1.0 or greater.
.»
• » BORLAnD
INTERNATIONAL
Available at Better Dealers Nationwide.
SideKick and SuperKey are trademarks of Borland Internallonal. Inc
IBM and PC· DOS are trademarks of International BUSiness Machines Corp
ProKey IS a trademark of RoseSoft
Telephone: (408) 438-8401
Telex: 172373
Telecopier: (408) 438-8691
For Credit Card Order Ca
(800) 255-8008,
(800) 742-1133 (in CA.)
WordStar IS a trademark of Mlcropro International Corp
NOTES
l1J'i7ln'\lij]'il~11( ilJ127
'~""l,~uili1'L"':\ '''~'::1)\'
.
,~ ./!-JL...::. i.-.JL-.4!-.4,-" '_ t __
INFOWORLD'S SOFTWARE PRODUCT OF THE YEAR

Just listen to what Dan Robinson has to say!
''SIDEKICK™ stands in the shadows behind whatever program you are using, ready to
jump forward when you need it. The program's various functions use windows that overlay
the display you are working with and restore the screen when you are through. The pro-
gram contains a respectable word processor for note taking, a dialer that your smart modem
can use with your phone list, a calculator for hexidecimal/binary/decimal arithmetic, an ap-
pointment calendar and a ASCII table for programmers ... SIDEKICK IS A TIME-SAVING,
WORK-SAVING, FRUSTRATION-SAVING BARGAIN·"_Dan Robinson of INFOWORLD
Here's SIDEKICK running over Lotus 1,2,3TM. In the SIDEKICK All the SIDEKICK windows stacked up over lotus 1,2.3. From bot-
Notepad you'lI notice data that's been imported directly from the tom to top: SIDEKICK's "Menu Window': ASCII table, Notepad,
Lotus screen. In the upper right you can see the Calculator. Calculator, Appointment Scheduler and Calendar, and Phone
Dialer. Whether you're running WordStarTM, Lotusn.t, dBasen.t, or
any other program, SIDEKICK puts all these desktop accessories
instantly at your fingertips.
THE CRITIC'S CHOICE
- Charles Petzold, PC MAGAZINE: "In a simple, beautiful implementation of Word-
Star'sTM block copy commands, SIDEKICK can transport all or any part of the display
screen (even an area overlaid by the notepad display) to the notepad."
- Garry Ray, PC WEEK: "SIDEKICK deserves a place in every PC."
- Jerry Pournelle, BYTE: "If you use a PC, get SIDEKICK. You'll soon become depen-
dent on it."
SIDEKICK is an unparalleled bargain at only $54.95 (copy-protected)*
$84.95 (not copy-protected)*
"Minimum System Configuration: SIDEKICK is available now for your IBM pc, >IT. AT. jr. and 100% compatible microcomputers. The
IBM PCjr. will only accept the SIDEKICK not copy-protected version. Your computer must have at least 128K RAM, one disk drive and
PC-DOS 1.0 or greater.
SIdeKick IS a trademark of Borland International. Inc aBase IS a trademark of ~shloo·Tale. IBM and PC·DOS are lfaoemarkS of Internal,onai BUSiness
Machones Corp. InfOMlrld IS a trademark 01 Popular ComputIng. Inc.. a substo.ary 01 CW Cor,municatlOlls. Inc loCus and 1.2.3. are trademarks of lotus
Dew!lopment Co~ WordStar IS a trademark of M.cropro fnletnalional Corp
-BORLAnD Available at Better Dealers Nationwide.

Telephone: (408) 438-8400
Telex: 172373
Telecopier: (408) 438-8696
_ INTERNATIONAL Scotts Valley, CA 95066 For Credit Card Order Call:
(800) 255-8008,
(800) 742-1133 (in CA.)
NOTES
TURBO PASCAL 3.0:
BUSTS THE COMPILER SPEED BARRIER!
Borland International, the creators of SIDEKICK™ and SUPERKEyTM, now launch TURBO PASCAL™ 3.0-The world's fastest compiler!
MS Pascal TURBO 3,0
Look at the Power: Compilation Speed:(")
(And twice as fast as TURBO 2.0!) A quick
206 sec 8 sec
Execution Speed:(")
(Resulting programs run faster. A real plus for TURBO PASCAL
programs over other compiled programs.) A speedy
20 sec 9 sec
Code Size:
(The same size as TURBO 2.0. How'd we do it?) A small
35K 12K
Complier Size:
(It hasn't changed from TURBO 2.0.) A mere
300K+ 35K
(.) Benchmark run on an IBM PC using MS Pascal version 3.2 and the DOS linker version 2.6. The 179 line program used is the "Gauss·
Seidel" program out of Alan R. Miller's book, Pascal Programs for Scientists and Engineers (Sybex, page 128) with a 3-dimensiona
non-singular matrix and a relaxation coefficient of 1.0.
TURBO PASCAL 2,0 had a lot of great features and they're stili there in TURBO PASCAL 3,0:
• BUILT-IN INTERACTIVE EDITOR: WordStar™-easy editing lets you debug quickly.
• ONE-STEP COMPILE: No more 'hunting & fishing' expeditions. TURBO PASCAL does it all in one simple step.
• MICROCALC: A full-blown spreadsheet on your disk ready-to-run with complete annotated source code.
The best Just gets better with a host of additional features:
• We rewrote the I/O system and we now support I/O redirection.
• For iBM PC versions we've added TURTLE GRAPHICS and full tree directory support.
• For all 16-bit versions, we now offer two options: 8087 math coprocessor sLipport for intensive calculations and Binary Coded
Decimals (BCD) for business applications.
Debugging is So Easyl When you run your program, TURBO 3,0 locates any error, takes you to It, and tells you what it
is, Your program Is up in record time,
Specifications:
Compile Speed Approximately 150 lines per Code size: Object code plus 9K for run time
second (IBM PC) library(complete)
Number of Max code size: 64K
passes: 1 Max data size: 64K
Complied to: Native machine code (.COM files Max stack size: 64K
I.CMD for CP/M-86)) Max heap: Unlimited
The Critics Unite:
Jeff Duntemann, PC MAGAZINE: "Language deal of the century ... TURBO PASCAL. It introduces a new programming environmenl
and runs like magic."
Dave Gariand, POPULAR COMPUTING: "Most Pascal compilers barely fit on a disk, but TURBO PASCAL packs an editor, compiler
linker, and run-time library into just 29K bytes of RAM."
Jerry Pournelle, BYTE: "What I think the computer industry is headed for: well-documented, standard, plenty of good features, anc
a reasonable price."
The unbelievable TURBO PASCAL 3.0 is available right away for $69.95 (not copy-protected)·,
• Options: TURBO PASCAL with 8087 or BCD at a low $109.90. TURBO PASCAL with both options (8087 and BCD) priced at $124.95.
Minimum System Configuration: TURBO PASCAL 3.0 is available today for your Z-80, 8088/86, 80186 or 80286 microprocessor com·
puter running either CP/M-80 2.2 or greater, CP/M-86 1.1 or greater, MS-DOS 1.0 or greater or PC-DOS 1.0 or greater. Your computel
must have at least 64K RAM and one disk drive. A XENIX version of TURBO PASCAL will soon be announced, and before the enc
of 1985, TURBO PASCAL will be running on most 68000-based microcomputers.
Integrated full screen editor with WordStar·like functions. Supports BIOs/BDOS calls for CP/M·80 and 86. Supports Interrupt and MS·DOS function calls for MS
DOS. PC· DOS. Inline machine code. External subroutines (in relocatabie binary format). Direct access to CPU and data ports.
Available at Better Dealers Nationwide, Telephone: (408) 438-8400
ID BORlAnD
INTERNATIONAL
Telex: 172373
Telecopier: (408) 438-8696
For Credit Card Order Call:
(800) 255-8008,
(800) 742-1133 (in CA,)
TurboPascal. SideKick and SuperKey are trademarks of Borland International. Inc. CP/M is a trademark of Digital ResearCh, Inc. IBM and PC· DOS are trademark!
of International Business Machines Corp. MS·DOS and XENIX are trademarks of Microsoft Corp. WordStar is a trademark of Micropro International Corp. Z·8e
is a trademark of Zilog. Inc.
NOTES
TURBO TUTOR
Learn Pascal from the folks who invented TURBO PASCAL™ and TURBO TOOLBOX™.
Borland International proudly introduces TURBO TUTOR™. The perfect complement to
your TURBO PASCAL compiler.
• TURBO TUTOR is really for everyone-even if you've never programmed before.
If you're already proficient, TURBO TUTOR can sharpen up the fine points.
The 300 page manual and program disk divides your study of Pascal into three learning
modules:
For the Novice: Gives you a concise history of Pascal, tells you how to write a simple pro-
gram, and defines the basic programming terms you need to know.
Programmer's Guide: The heart of TURBO PASCAL. This section covers the fine points
of every aspect of TURBO PASCAL programming: program structure, data types, control
structures, procedures and functions, scalar types, arrays, strings, pointers, sets, files and
records.
Advanced Concepts: If you're an expert, you'll love the sections detailing subjects such
as "how to use assembly language routines with your TURBO PASCAL programs."
A MUST. You'll find the source code for all the examples in the book on the accompanying
disk ready to compile.
TURBO TUTOR may be the only reference on Pascal and programming you'll ever need!
TURBO TUTOR is a real education for only $34.95. (not copy-protected).*
"Minimum System Configuration: TURBO TUTOR is available today for your computer running TURBO PASCAL for PC-DOS, MS-DOS,
CP/M-8a. and CP/M-86. Your computer must have at least 128K RAM, one disk drive and PC-DOS 1.0 or greater, MS-DOS 1.0 or greater,
CP/M-80 2.2 or greater, or CP/M-86 1.1 or greater.
• » BORLAnD
. ) INTERNATIONAL
Available at BeHer Dealers Nationwide.
Telephone: (408) 438-8400
Telex: 172373
Telecopier: (408) 438-8696
Scotts Valley, CA 95066 For Credit Card Order Call:
(800) 255-8008,
(800) 742-1133 (in CA.)
Turbo Pascal. Turbo Tutor and Turbo Toolbox are trademarks 01 Borland International. Inc. CP/M is a trademark of D,gital Research. Inc. MS·DOS is a trademarl
of Microsoft Corp. PC-DOS is a trademark of International Business Machines Corp.
NOTES
TURBO TOOLBOX
"A PASCAL PROGRAMMER'S BEST FRIEND"
Borland International now offers TURBO TOOLBOXTv , a three-module TURBO package that gives
even a novice programmer the expert's edge. The perfect complement to TURBO PASCALrM , the in-
dustry standard in Pascal compilers. Fully commented source code to the first two modules is includ-
ed on diskette.
Take a Close Look:
TURBO-ISAM Files Using B+ Trees
A powerful implementation of the state-of-the-art B+ tree ISAM technique. The TURBO-ISAM
system makes it possible to access records in a file using keys (e.g., "Smith" or "Rear Bumper")
instead of just a number. Even though you have direct access to individual records in the file,
you also have easy access to the records in a sorted sequence. A must if you plan to work with
direct access files. Source code is included in the manual.
QUICKSORT On Disk
The fastest way to sort your data. A super efficient implementation of the quickest data sorting
algorithm preferred by knowledgeable professionals. Available for you now with commented source
code.
GINST (General Installation Program)
Now, your programs are easily installed on other terminals. Saves hours of work and research,
and adds tremendous value to everything you write.
Free Sample Database:
Your TURBO TOOLBOX now comes with source code for a working database... right on the disk.
Just compile it, and it's ready to go to work for you. It's a great example of how to use TURBO
ISAM, and at the same time, it's a working piece of software you can use right away. The database
searches by name or code, updates records as needed. Since you're supplied with commented
source code, you can tailor it to your own specific needs.
Specifications:
Random Record Access with ISAM files using B+ trees. Turbo Access Speed: 1.5 seconds (Average seek time with 1.000 records-86
bytes/rec). Quicksort speed: 43 seconds (1.000 records. 86 bytes/rec.. 256K RAM).
Jeff Duntemann, PC Tech Journal: .. Turbo Pascal is the most nearly bug-free compiler I have eVI
used, and the Turbo Toolbox comes close in quality. "
Jerry Pournelle, BYTE: "The tools include a B+ tree search and a sorting system; I've seen stuft
like this, but not as well thought out, sell for hundreds of dollars."
TURBO TOOLBOX is only $54.95 (not copy-protected):
'Mlnimum System Configuration: TURBO TOOLBOX is available today for your computer running TURBO PASCAL 2.0 or greater fo
PC· DOS. MS·DOS. and CP/M·SO or Turbo Pascal 2.1 or greater for CP/M·86. Your computer must have at least 12SK RAM. one disl
drive and PC·DOS 1.0 or greater. MS·DOS 1.0 or greater. CP/M·SO 2.2 or greater. or CP/M·86 1.1 or greater.
I DBORlAnD Available at BeUer Dealers Nationwide. Telephone: (408) 438-840

4585 Scotts Valley Drive Telex: 172373
DINTERNATIONAL Scotts Valley, CA 95066
Telecopier: (408) 438-869
For Credit Card Order C~
(800) 255-8008,
Turtlo Pascal and Turco ToolCox are lrademar~s 01 Borland Internat,onal. Inc
CP/M ,s a trademarl< 01 Digital Research. Inc (800) 742-1133 (in CA.)
MS·DOS IS a trademarl< of M,crosoft Corp
PC· DOS 's a trademarl< of International Bus,ness Mach,nes Corp
NOTES
TURBO GRAPHIX TOOLBOX
HIGH RESOWTION MONOCHROME GRAPHICS
FOR THE IBM PC AND THE ZENITH 100 COMPUTERS
Dazzling graphics and painless windows. The Turbo Graphix Toolbox™ will give even a
beginning programmer the expert's edge. It's a complete library of Pascal procedures that
include:
• Full graphics window management.
• Tools that allow you to draw and hatch pie charts, bar charts, circles, rectangles and a
full range of geometric shapes.
• Procedures that save and restore graphic images to and from disk.
• Functions that allow you to precisely plot curves.
• Tools that allow you to create animation or solve those difficult curve fitting problems.
• and much, much more ....
No sweat and no royalties. You can incorporate part, or all of these tools in your programs,
and yet, we won't charge you any royalties. Best of all, these functions and procedures come
complete with commented source code on disk ready to compile!
John Markoff & Paul Freiberger, syndicated columnists: "While most people only talk
about low-cost personal computer software, Borland has been doing something about it.
And Borland provides good technical support as part of the price."
The versatile TURBO GRAPH IX TOOLBOX is priced at only $54.95 (not copy-protected).·
"Minimum System Configuration: Turbo Graphix Toolbox is available today for your computer running Turbo Pascal 1.0 for PC-DOS.
Ms-DOS. and CP/M-86. Your computer must have at least 128K RAM. one disk drive and PC-OOS 1.0 or greater. MS-DOS 1.0 or greater.
or CP~ 1.1 or greater.
Available at Better Dealers Nationwide.

-BORlAnD 4585 Scotts Valley Drive
Telephone: (408) 438-8400
Telex: 172373
Telecopier: (408) 438-8696
_ INTERNATIONAL Scotts Valley, CA 95066 For Credit Card Order Call:
(800) 255-8008,
Turbo Pascal and Turbo Graphix Toolbox are trademarks 01 Borland Intemalional. Inc.
CPIM is a trademark 01 Digital Research. Inc. (800) 742-1133 (in CA.)
IBM and PC-DOS are trademarks 01 Intemational Business Machines Corp.
MS-DOS is a trademark 01 Microsoft Corp.
Zenith is a trademark 01 Zenith Radio Corp.
NOTES
Wbetber for buIDea pmfeuioJWs.lOftwue eaaJDeera. students or computer bobbytata.
IorIud offen I better IOlutiCl!l:
Slde~ck ,TW a combination notepad. calculator. appoint. Thrbo Pilsul ThoibOll:,TW a set of three fundamental
ment calendar. autodialer and ASCII reference table that utilities that work In conjunction with Thrbo PascaL Thrbo-
can be used With any application. without exiting or ISAM flles using B+ trees ; QUicksort on disk; and the
rebooting GINST (General InstallatlOnl program. which allows pro-
SuperKey,TIoI a unique ProKey'" compatible keyboard grams wntten In Thrbo Pascal to have terminal installation
utility that proVides macros to reduce cumbersome modules. Complete annotated source code is included on
keyboard commands to Single keystrokes of any choosing. disk and no royalties are required If you include parts or all
plus a powerful encryption tool to prevent unwanted entry of the Toolbox In your programs.
Into confidential computer files. Thrbo Graphix Toolbox ,TW high resolution graplucs for the
Thrbo Pascal,''' the micro Industry's newest standard In IBM!> PC family of computers and the Heath 2-100: com-
Pascal compliers, IS now avallable With optional 8087 plete with windowing procedures. As with the Thrbo
support for 16·blt computers. Thrbo Pascal for the Apple Pascal Toolbox. you get full commented source code on
Macintosh and other 68000 systems will soon be available disk. and pay no royalties for programs you write incor-
as well. porating Thrbo Graphix procedures.
Thrbo Pascal Thtor,''' a step-by-step tutorial on program-
BORLAND INTERNATIONAL. ming and Turbo Pascal. complete With commented source
code for all program examples on disk.
INTERNATIONAL
4)8~ Scoos Valley Drive
Scoos Valley, California 9)OS6
(4a!) 438-8400
Look for Borland software at the computer dealer near you or for VISA and MasterCard orders
call toll free , I (800) 255·8008: In California 1 (800) 742·11H. Lines are open 24 hours a day. 7 days a week.

Turbo Tutor Version 1.0 Jan85

Uploaded by

Copyright:

Available Formats

Turbo Tutor Version 1.0 Jan85

Uploaded by

Document Information

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

Turbo Tutor Version 1.0 Jan85

Uploaded by

Copyright:

Available Formats

The TURBO

This software is protected by both United States Copyright Law and

BORLAND INTERNATIONAL, INC., SPECIFICALLY DISCLAIMS ALL OTHER

Third Edition. January. 1985

PART I TURBO PASCAL FOR THE ABSOLUTE NOVICE

PART II A PROGRAMMER'S GUIDE TO TURBO PASCAL

PART III ADVANCED TOPICS IN TURBO PASCAL

language, but probably didn't want to pay the high prices

real treasure chest of information when you know more about

Since TURBO Pascal is available for so many different types of

Although TURBO Pascal will find most of your syntactical errors

2. A SHORT HISTORY OF PASCAL

UIII Iplll ifill! 11l{

-An Early Programming Attempt-

A programming language is nothing more than a program

Both these languages, COBOL and FORTRAN, did what they

of TURBO Pascal's unique characteristics is that it is very small,

3. WRITING A SIMPLE PROGRAM

Did you know that computers can show up a basic difference

Also 'you should remember that as a human being you have

In other words, with computers, as with no other part of life, it will

The compiler uses this information to know where all the

The compiler uses the definition statements as tools when it

Third: the program statements themselves:

The program statements tell the compiler specific steps the

4. PROGRAMMING IS MORE THAN

The process of programming a computer is not really about

-Just Sitting Back Thlnklng-

Computers are pretty consistent machines. If they do what you

Someti mes it seems that no matter where we go, we are

If it is a simple decision as to whether or not to have a car, your

As a practical matter, we divide the data to be stored in a

In orderto store data items in a computer, we need places-similar

When we want to convey to one another one thought, or a

By this time, I hope you are getting the impression that a

-Planning, the First Step-

The best way to go about developing a program is to do so

put all of them together, the large problem will be solved. In

By the way, this method of program design is ideally suited to

One general note:While you are thinking about the problem to be

The only reason to store data in your computer is to get it back

The best way to go about it is to fully describe each piece of data.

As another example, let's say that your program is going to ask

This is where you put down reminders to yourself about things

Do you remember the program we ran back in Chapter 3? Let's

Miscellaneous Stuff: I want my name to appear 20 times; I'll use a

1. Start the program

4. Erase the computer's screen.

Sometimes it seems that the hardest part of doing anything is

6.1 TURBO PASCAL

To start your adventure in good programming (did he really say

6.2 THE MAIN MENU

When we say menu we usually mean a Iist of items from which we

This shows the path name of the current default directory. To

to indicate the Active Directory option.

- TURBO Pascal Edltor-

What is happening is this: If you have specified the name of a

- The TURBO Pascal Compiler-

When you press C, TURBO Pascal takes your edited (source

The TURBO Pascal compiler compiles programs, and then puts