Web and Multimedia Systems Development – State of the Practice

Dijana Plantak Vukovac

University of Zagreb
Faculty of Organization and Informatics, Varaždin
[email protected]

Božidar Kliček
University of Zagreb
Faculty of Organization and Informatics, Varaždin
[email protected]

Abstract. Over the past few years there has been manage requirements, how to control development
much academic interest in the development of com- processes, how to co-ordinate the collaborative work
plex hypermedia systems, particularly web-based of design teams, and how to effectively manage pro-
ones. The literature suggests that the dominant devel- jects" [2].
opment approach is the ad hoc approach, which fails Some authors [13, 14, 17, 20] suggest that the
to address many aspects of hypermedia systems de- dominant hypermedia development approach is an ad
velopment. Hence, many specialized web/hypermedia hoc approach with "implement, test, release" devel-
design methods have been introduced to discipline the opment steps, which fails to address many important
development process and resolve issues typically not aspects of complex hypermedia systems development.
covered by software engineering development para- Furthermore, "most web applications are developed
digms. However, some studies have shown that web with little or no control and without a formalized
and multimedia development is not as chaotic as it is process" [22]. So, whereas on the one hand, there is a
generally considered despite the fact that the afore- lack of a systematic approach and appropriate meth-
mentioned specialized methods are not widely used. odologies, on the other hand important issues are ig-
In order to explore Croatian web design organiza- nored because the current software engineering
tions' practice of web and multimedia development, a development paradigms do not fully cover the exist-
web survey was conducted. In this paper the study ing hypermedia features of web and multimedia sys-
findings are presented. tems.
Although the ad hoc approach is acceptable for
Keywords. Multimedia, Web-based Systems, Web creating simple web sites, it fails to deliver complex
Development Practice web-based systems of satisfactory quality [13, 17].
Thus a Cutter Consortium survey, referred to by
Ginige in [7], showed that the majority of delivered
1. Introduction systems neither demonstrated required functionality
nor met business needs, whereof about half of them
With the advancement of numerous enabling tech- were of poor quality. Furthermore, the development
nologies, particularly web-based ones, multimedia us- process was behind the schedule and exceeded the
age is moving from offline media to the online budget estimates.
medium. This shift has been accompanied by the in- This leads to the conclusion that large web-based
troduction of different kinds of applications ranging systems development is experiencing the same prob-
from simple online multimedia presentations (e.g. lems inherent in the IS development. However, hy-
brochureware web sites) to large-scale hypermedia permedia features inevitably give rise to a number of
applications offering rich multimedia content and new issues, which makes web-based systems devel-
multiple navigational paths (e.g. complex web sites of opment different from IS development [16, 17, 22]. It
the Olympic Games or World Championships). Also, is the hypermedia features, defined as different types
multimedia and web technologies have been inte- of media connected by hyperlinks, allowing users to
grated into business information systems to support navigate through an information space following
organizational activities such as sales, marketing, various paths, that create the point of distinction be-
support or collaboration, on a B2B and/or a B2C tween the development of multimedia and web-based
level. Since web sites that facilitate business proc- systems on the one hand and information systems on
esses have more in common with traditional cli- the other. In this distinction several aspects are im-
ent/server applications than with static web sites, the plied [14, 16, 17, 18, 20]:
software engineering principles are essential in the • non-linearity of hypermedia documents facilitates
development of such systems [20]. the occurrence of a complex structure subse-
Thus, "issues similar to those encountered in tradi- quently hard to maintain as the application size
tional IS development have emerged – such as how to grows;
• aesthetic and cognitive aspects of web/ in popular IT magazines. Another reason may be the
hypermedia systems are more important than in questionable practical value of some of the methods
traditional systems; which neither prescribe the whole system life cycle
• the audience of a web site is fairly wide, with us- nor have flexible design tool support.
ers having different information needs and com-
puter skills, so usability and accessibility aspects 2. Review of previous research
are of great importance;
• different development activities imply the need Besides the web/hypermedia-specific methods us-
for a team with different professional back- age, the overall web/hypermedia development prac-
grounds: graphic designers, programmers/ tice has been a topic of interest to many researchers.
developers, systems analysts, information archi- In 1998 Vora [26] conducted voluntary mixed-
tects, technical writers, media producers, project mode survey among web professionals in the USA to
manager(s), marketing specialists etc.; explore the characteristics of the web design process.
• rapidly changing web environment forces shorter He found that designers used design guidelines or
life cycles of web/hypermedia systems; style guides while developing web pages, which they
• continuous changes in information content, tech- subsequently optimized and evaluated, but were un-
nologies and standards used imply that develop- happy with browser incompatibilities and insufficient
ment should be incremental and iterative. web sites usability.
Russo & Graham [22] conducted a survey among
As a result, drawing on the fact that the range of web developers of five hundred web sites indexed in
hypermedia applications is very broad indeed, over the 1997 Internet & Web Yellow Pages. Having ob-
fifty new methodologies, methods, techniques and tained the response rate of just over 10%, they con-
models1 for web/hypermedia design have been pro- cluded that the majority of respondents did not have
posed in the last decade or so, to respond to the call web development standards or guidelines and none of
for "sound methodology, disciplined and repeatable them was using a formal development methodology.
process, better development tools, and a set of good A minority of developers, though, was using tradi-
guidelines" in order to "bring the current chaos in tional development tools and diagramming tech-
Web-based system development under control..." [7]. niques, such as ER diagrams, flowcharts etc.
In most of these methods the Design phase, com- Taylor et al. [24] conducted multiple case studies
prising Conceptual design, Navigational design and about activities, techniques and standards in web site
Presentational design, is prescribed in detail. Some development in different organizations in the North
of the methods cover all the phases of a West of England. They found that the ad hoc ap-
web/hypermedia system life cycle, from Require- proach was the dominant development approach, spe-
ments capture & Analysis to Testing & Maintenance. cialized techniques were scarcely used and some of
The methods are based on the following software en- the phases in the web site life cycle had been skipped.
gineering techniques: entity-relationship techniques Half of the web designers had followed some formal
(methods such as Hypertext Design Method, HDM guidelines for the web site development but they had
[6]; or Relationship Management Methodology, been developed in-house and had nothing in common
RMM [8]), object-oriented techniques (methods such with the recognized web standards.
as Object-Oriented Hypermedia Design Method, In Australia, Venable & Lim [25] conducted a
OOHDM [23]; Hypermedia Flexible Process Model- postal survey about web information system (WIS)
ing, HFPM [9]; UML-based Web Engineering, UWE development methodologies. With the response rate
[15]), or a combination of techniques (methods such of only 9.32%, they found that 66.7% of respondents
as Web Modeling Language, WebML [4]; World were using some form of methodology (mostly em-
Wide Web Design Technique, W3DT [3]; Web Site bedded in the web development tools) to guide their
Design Method, WSDM [5]). WIS development, while RMM and W3DT were the
However, despite the fact that many research pa- only web/hypermedia-specific methods tried and
pers about web/hypermedia-specific methods have used.
been published, a few studies [1, 10, 25] conducted to Lang [10, 11] conducted dual-mode survey among
explore the usage of those methods in practice Irish hypermedia developers in 434 organizations who
showed that most practitioners were not familiar with specialized in web/hypermedia systems development
them and did not use any web/hypermedia-specific or whose companies had IS departments. This study is
methods. One of the reasons for that might be insuffi- recognized as the most comprehensive study of hy-
cient dissemination of the methods at universities and permedia systems development lately. With the re-
markable response rate of 44.5%, Lang found that
1
Methodologies, methods, techniques and models are terms widely most organizations were using a formalized or unfor-
used to name web/hypermedia design methods (methodologies). malized development process that had clearly defined
As these terms are often vaguely defined, they can either have a tasks or phases in it, wherein 23% of respondents
different or a similar meaning for the authors of methods (meth-
were using a hybrid, customized or proprietary in-
odologies). Therefore, to avoid broad nomination, the term
method is used in this paper. house method or approach [12]. More than two thirds
of the respondents indicated that they had used for- marketing, branding, advertising, telecommunica-
malized procedures or guidelines in hypermedia sys- tions, web hosting;
tems development. • portfolio and company information, including e-
Zhou & Stålhane [27] explored web system devel- mail address (or contact form), is available on the
opment practice in 11 ICT organizations in Norway. company’s web site.
They found that the majority of respondents (80%)
were using the iterative and incremental development The population sample was compiled from five
approach, but half of them did not have a process different business search engines (Croatian Company
model at all. Moreover, the level of the usage of engi- Directory issued by the Croatian Chamber of Econ-
neering methods and techniques for reliability and ro- omy, WLW Business Search Engine, T-Portal Web
bustness was not high. Directory, VIP Web Directory and Net.hr Web Direc-
Rosson et al. [21] conducted a survey among in- tory) and included companies from all the 21 Croatian
formal and experienced web developers in the USA. counties.
E-mail invitations to participate in the web survey The final population sample was filtered for dupli-
were sent to 591 organizations, 334 (55%) of which cates, with the sample eventually consisting of 418
sent back their responses, with 55% of the responses companies. All of them were sent an e-mail with the
coming from web professionals. They found that web research description, a unique link to the web survey
development process did not begin with the web page for access control and authorization data (username
construction, but rather with some prior planning. The and password). The web survey was available online
major problem identified by the survey was unsatis- for 31 days and during that period, after two remind-
factory collaboration within the web team. ers about the survey, 169 responses by design-
Although researches mentioned above were con- ers/developers were received, which amounts to the
ducted throughout the world, no similar study had response rate of 40.43%. However, 68 surveys were
been performed in Croatia. It is based on this very fact insufficiently completed, so the usable response rate,
and by taking into consideration all the relevant issues based on the true population size, was 23.71%, or 101
in current web/hypermedia development that the au- participants.
thors of this study set the following research goals:
a) to gain a broader view of web and multimedia 4. Research results
systems development practice in Croatia (design-
ers/developers profile, web and multi-media pro-
4.1. Designers/developers profile
jects characteristics);
b) to identify the exact way in which designers per-
The survey data analysis revealed that an average
form the development process as well as methods
Croatian web or multimedia designer/developer is a
and techniques they use in design;
younger male person (87.13% of respondents were
c) to explore opinions and trends in web develop-
male, average age 32) working in a small sized com-
ment.
pany employing up to ten employees (80.20% of re-
spondents).
3. Research methodology The majority of respondents graduated from high
school (39.60%) or university (54.45%) while just a
The study was based on Lang's questionnaire [9], minority of them holds a scientific degree (3.96%).
subsequently adapted to suit the local terminology and Their professional backgrounds vary significantly.
modified with questions concerning phases and activi- Some of those who finished secondary school identi-
ties in the development process, as well as opinions fied themselves as having a high-school-diploma, or,
about web development trends. The research method more precisely, as electronics engineers, graphics
used was a web survey, with the majority of questions technicians, economists, and even aircraft mechanics,
being close-ended, and a few open-ended questions nautical-school students, construction workers etc.
with a blank field for providing additional comments Those who finished tertiary education identified
to the survey. themselves as bachelors of informatics, electrical en-
The population of interest for this research had to gineers, graphics engineers, Bachelors of Economics,
satisfy the following criteria: and even high-school teachers, Bachelors of Journal-
• the company is registered in a court register and ism, architects, aviators, etc. They perform different
indexed in any one of the five online business design roles and functions, including graphic design-
registers; ers, web designers, web programmers, developers,
• developing web and multimedia systems can be webmasters, multimedia programmers, project man-
broadly classified as one of the company’s pri- agers, art directors etc.
mary business activities; this may include some Additionally, in the survey the respondents
of the following: web site development, web ap- claimed to have an advanced or very advanced level
plication/system development, software and/or of proficiency in different professional disciplines
information system development, multimedia relevant to web/hypermedia systems development:
system development, graphic and visual design, computer programming/software engineering, graphic
design, information architecture, user interface design developed by 38.39% respondents. That is similar to
and technical writing. The only area of which more Vora's findings [26] but quite different from Lang's
than half of respondents have only basic knowledge, findings [10], where two-thirds of respondents stated
or no knowledge at all, is film production, which is that the size of the systems developed exceeded 50
consistent with the findings that almost half of re- pages/screens.
spondents’ companies are outsourcing audio/video The average Croatian web team have 3 design-
producers (as 49.40% of 79 respondents claimed) and ers/developers (mean = 3.34, median = 3), which is
specialists in animations (as 45.42% of 75 respon- consistent with the findings of other studies [10, 26,
dents claimed). 27]. One-man team is found in 13.79% of the compa-
Regarding their experience in developing different nies, two- or three-member teams are found in
types of web/hypermedia systems, the respondents 51.72% of the companies, four- to five-member teams
were able to claim it by specifying the number of are found in 26.44% of the companies, while only
years in their profession or the number of projects 8.05% of the companies have more than five web
they participated in. More than half of respondents team members.
(58.59%) had between 5 and 10 years’ experience, so The participants also had to indicate the actual and
on average, a designer/developer had about 7 years’ anticipated project duration and cost of their most re-
experience (mean = 7.13, median = 7). Immense cently completed large-scale project. The project du-
variations in the number of projects were indicated, ration (in weeks) was indicated by 71.74% of 92
ranging between 2 and 500, with almost 80% of re- respondents, while the project cost was acknowledged
spondents who participated in 20 or more projects. by only 38.37% of 86 participants who responded to
the questions (other possible answers were „Don't
4.2. Characteristics of web and multimedia know" or „Not planned/recorded").
projects To exclude extreme values gathered for both pro-
ject delivery times and costs, a 5% trimmed mean and
median were used as measures of central tendency. It
In the survey different characteristics of web and was found that the average planned delivery time was
multimedia projects were explored: types and features 8 weeks (5% trimmed mean = 7.81, median = 5.5),
of the developed systems and their complexity, as while the actual project duration was 11 weeks (5%
well as project management (team size, project dura- trimmed mean = 10.98, median = 6) or 2 and a half
tion and cost, documentation management). months (see Table 1). That is 3 weeks less than the
The most common type of web/hypermedia sys- average project delivery in Ireland [10].
tems developed was a simple brochureware web site:
88.30% of respondents had developed it during the Table 1. Project duration (anticipated and
past two years. The second most commonly devel-
actual)
oped type of system is the Content Management Sys-
tems (CMS), developed by 84.37% of respondents.
Anticipated Actual
More complex systems are also developed but less of-
duration duration
ten (up to 5 projects in two years): Web portals/Web
(N=66) (N=72)
directories, electronic catalogues, transactional busi-
ness applications, CD-ROM/DVD multimedia appli- 1-4 weeks 45.45.% 30.56%
cations. However, a vast majority of respondents had 5-8 weeks 30.30% 30.56%
never developed e-learning applications (74.74%), 9-12 weeks 6.06% 15.28%
news and other information services (61.29%) or 13-16 weeks 4.55% 6.49%
complex cross-referenced documentation (63.44%). A 17-20 weeks 1.52% 1.39%
few respondents added feed aggregators, web ser- 21-24 weeks 3.03% 1.39%
vices, blog systems, touch screen applications and in-
More than 24 weeks 9.09% 13.89%
teractive maps as web/hypermedia systems developed
in the past two years.
The features that characterize most commonly de- Comparing the planned and actual delivery times
veloped web/hypermedia systems include a database- of particular projects, it was found that 60.61% of
driven web site, dynamically generated web pages as projects exceeded the planned schedule, 21.21% of
well as frequent and significant changes in content. projects were delivered on time and 18.18% of pro-
However, many respondents claimed to have no ex- jects ended ahead of schedule.
perience developing systems with support for multiple The average planned cost of a complex project
delivery devices and web 2.0 applications (44.09% ranged from about 15,000 HRK, or approximately
and 43.01%, respectively) in the past two years. 2,000 € (median value), to 40,085 HRK, or approxi-
Considering the system complexity, expressed by mately 5,400 € (5% trimmed mean value), while the
number of screens/on-line pages [10], almost half of average actual costs were slightly higher (median ≈
respondents (46.46%) indicated that they usually de- 2,120 €, 5% trimmed mean ≈ 5,800 €), but much
veloped small web/hypermedia systems (up to 50 on- lower than in Ireland (median ≈ 18,000 €, 5%
line pages/screens) while large-scale systems were trimmed mean ≈ 41,382 € [10]). However, project
costs are far better managed than project schedules, ent. Many respondents indicated the need for clients'
with one-third of the projects being delivered within feedback after every phase, which suggests that the
the agreed costs and one-third of the projects experi- development process is iterative and incremental, as it
encing cost over-runs. was recognized in [27]. The practice development
To identify documentation management in pro- process is as follows:
jects, the participants were asked to indicate the num- • Requirements Capture & Analysis – activities in
ber of requirements and/or the number of pages this phase are:
within written requirements specifications, as well as o identification of purpose, goals, informa-
the usage of documented procedures or guidelines for tion needs, content and potential users of
web/hypermedia development. Out of 94 respondents, the application during the initial meeting
51.06% of them had written requirements specifica- with the client;
tions for the most recently completed project. Using o analysis of business processes, definition
documented procedures or guidelines for web/ of required functionality and selection of
hypermedia development is a common practice in a web technology and development tools;
minority of companies, as only 37.23% of 94 respon- o analysis of client's organization brand,
dents responded affirmatively, which is a percentage analysis of competitors’ web sites and
similar to Russo & Graham's findings [22] but much marketing research if necessary;
lower than in Lang's findings [11]. In most companies • Initial Proposal – the phase comprises:
procedures for project planning, interface design o setting the rough concept of the web site's
/usability, requirements documentation and technical graphic and navigational design;
design documentation were developed. o presenting a proposal to the client by
Despite the low percentage of using documenta- mock-ups and storyboards delivered on the
tion identified, Croatian designers/developers believe paper or in graphic editing software;
that project plans and working methods are essential o after approval, collecting the content (text
and have to be explicitly documented (94.70% and and other media) for the web site;
59.38% of respondents believe that, respectively). • Design Phase – in this phase three activities can
The lack of adequate design documentation repre- be distinguished:
sents a minor to moderately problematic aspect in the o Conceptual Design – includes the design of
web/hypermedia development practice. The most sig- the information architecture of the applica-
nificant problem concerning web project management tion, where the structure of the database is
is communication with clients, due to their volatile planned and different diagrams are built;
and changing requirements, followed by difficulties o Navigational Design – includes the con-
preparing accurate time and cost estimates as well as struction of models with different naviga-
those controlling project scope and feature creep. On tional paths;
the other hand, the design of the user interface or lack o User Interface Design – several templates
of guidance in the use of design methods and tech- of user interface are sketched and authored
niques were not perceived as a problem by most of in HTML/CSS. At this stage different mul-
the respondents. timedia elements are edited and optimized;
• Development or Programming Phase (Implemen-
4.3. Development process tation) – activities in this phase are:
o database construction, programming mod-
Our study is in line with the international studies ules development and/or CMS design or
[1, 11] stating that web development is not as undis- upgrade;
ciplined as it is generally considered: 85.15% of re- o system integration;
spondents indicated that their organization used a o content loading;
web/hypermedia development process that had clear • Testing & Implementation of Subprojects – dif-
tasks and/or phases. In 62.38% of cases the process ferent types of testing are carried out as well as
used was not explicitly documented, while in 22.77% the implementation of smaller subprojects. Ap-
of cases the process was explicitly documented. plication is also upgraded on the basis of user
To further explore the web/hypermedia develop- evaluation or new specific requirements;
ment process in practice, participants were given an • Launching – a "soft" launch of beta version is
open-ended question to describe phases and/or activi- performed before client approval and final re-
ties of their organization's development process. Fifty- lease;
one respondents (50.49%) provided a description of • Education & User documentation – if needed,
the development process in their organization. education is performed to introduce client or
After the analysis and classification of all the re- maintenance team with the system functions.
sponses, several phases and activities were identified User and technical documentation is also pro-
in the development process in practice. The phases vided;
vary according to the purpose and complexity of the • Release – the phase comprises:
future web-based system and the contract with the cli-
 o a "hard" launch when the application goes centage compared with the findings of two other
online for the wide audience; studies [11, 25] which identified 4 and 2 respondents,
o other activities such as site promotion and respectively, having ever used some of those methods.
search engines optimization; As far as the authors of this study are aware, no
• Maintenance – the last phase comprises mainte- web/hypermedia development methods have been in-
nance of the system and its enhancements. troduced by Croatian universities yet. We believe that
such a high percentage represents a bias resulting
When comparing this practice development proc- from the respondents' tendency to rather provide a
ess with the development phases of specialized positive than a negative answer and the fact that those
web/hypermedia design methods, it is evident that it methods were first on the list of the proposed method
consists of all phases identified in the design methods categories. These findings imply the need for more
and described in [14], [16] and [22]. The practice de- extensive research to explore whether and how these
velopment process goes even further to include web methods are used in practice. Also, regardless of
project management activities. What is different, whether design methods are used or not, it would be
however, is the terminology used for naming the worthwhile to further examine the development proc-
phases in practice and literature. In the literature the ess approaches.
term "implementation" is used for "application build- Consistent with the findings indicating a relatively
ing", but Croatian developers denote application low percentage of use of proprietary software devel-
building as the Development or Programming Phase opment methods and approaches, it was also found
and broadly use the term "implementation" to denote that formalized software diagramming techniques for
the Launching Phase (similarly, "implementation" modeling the conceptual architecture of web/
considered as "installation of web application" is hypermedia systems, such as entity-relationship dia-
found in [22]). Another difference identified is a very grams, class diagrams, statecharts and use case dia-
loose margin between the Requirements Capture and grams, have a significantly lower percentage of use
the Analysis phases, so in the practice process devel- (between 42.9% and 22.2%) than informal techniques
opment they are comprised within one phase. (e.g. storyboards, 79.6%) and semi-formal techniques
To identify whether the web/hypermedia devel- (e.g. flowcharts, also 79.6%). Moreover, 36.5% of 96
opment process in practice is supported by any of the respondents believe that diagramming techniques are
methods, a closed multiple-choice question was intro- not vital to communicate the conceptual structure of a
duced. More than half of participants, or 52.48% out web/hypermedia system (in contrast to 4.9% of Irish
of 101 respondents, indicated that they used in-house respondents [10]).
or hybrid methods. Only 5.94% of respondents used
proprietary methodology, while 26.73% of respon- 4.4. Opinions about web/hypermedia systems
dents did not know or 14.85% of them did not use any development
methods.
To further explore the usage of different methods In the study, opinions about web/hypermedia sys-
and approaches in web/hypermedia systems design, a tems development were also explored. In Table 2 the
list of them was introduced, based on the categories respondents' attitudes to the development guidelines
proposed by Lang in [10]. The participants were thus and web trends are presented.
asked to indicate the extent of use of these methods in Interestingly, 30.21% of respondents believe that
their organization's development process. Not surpris- the Web 2.0 concept is more a matter of current trends
ingly, the majority of respondents often or continu- in the web development and less a result of actual
ously used tool-driven development approaches, e.g. user needs, to which 27.08% of respondents disagree.
PHP, Java, Flash, ASP, J2EE (70.33% of 91 respon- Whereas 20.83% of respondents are neutral in that re-
dents). The other top response category was object- spect, 21.88% of respondents do not know or have no
oriented development methods and approaches, e.g. opinion about Web 2.0, which leads us to the conclu-
OOA&D, UML and MSF (25.29% of 87 respondents), sion that a significant number of respondents are not
followed by rapid or agile development methods and familiar with the new ways of using the Web as a
approaches, e.g. RAD, Extreme Programming platform. Also, this lack of knowledge upholds the
(19.54% of 87 respondents). The majority of respon- finding that respondents have difficulties following
dents did not know or use any of the methods or ap- current development trends, which was recognized as
proaches (with the exception of tool-driven a minor to moderate problem by more than half of re-
development approaches) and the top three least spondents (54.25%).
known categories were specialized web/hypermedia However, the majority of respondents (97.89%)
development methods, traditional "legacy" software indicated the internet as the most useful informal
development methods and approaches, and incre- knowledge source. Services as varied as forums, por-
mental or evolutionary methods and approaches. tals, mailing lists and Microsoft Developer Network
Surprisingly, 12 respondents (13.48%) indicated (MSDN) for specialized subjects were specified as
that they used specialized web/hypermedia develop- means of acquiring and sharing information. Other
ment methods to some extent, which is a high per- very useful information sources indicated are observ-
ing or consulting experienced colleagues (94.79%), useful (40.63%) and not useful (38.54%). Possible
reading books, periodicals and magazines (86.46%), reasons for the negative attitude are inconsistent and
and consulting external specialists (77.9%). Many re- out-dated curricula and rapid advancements in web
spondents (30.21%) are neutral about the usefulness technologies, so "workshops about usability, naming
of organizational policies and procedures with a slight conventions, formalized development methods, team
tendency towards the positive attitude, but the same management and documentation management, would
percentage of them do not use this source of knowl- be very useful", as one respondent commented at the
edge at all. About the same percentage of respondents end of the survey.
consider formal education (secondary or tertiary) both

Table 2. Attitudes to development guidelines and web design trends

Don't know /
N Disagree Neutral Agree
No opinion
Client requirements are more important than user needs and
95 1.1% 59.0% 23.2% 16.8%
satisfaction.
Visual design is more important than the content and func-
98 0.0% 67.4% 26.5% 6.1%
tionality of the web/hypermedia system.
Navigation should be consistent and uniform throughout the
97 1.0% 7.2% 4.1% 87.6%
web/hypermedia system.
Usability and accessibility guidelines increase the quality of
94 5.3% 2.1% 11.7% 80.9%
a web site.
Development of the web site according to W3C standards
and testing decreases maintenance costs and facilitates en- 96 10.4% 17.7% 14.6% 57.3%
hancements.
Web 2.0 applications are more a matter of new trends and
96 21.9% 27.1% 20.8% 30.2%
less a matter of justified user needs.

5. Conclusion holders in building better web and multimedia sys-

tems have been identified by the authors:
The findings of this research reflect the web/ • conducting new qualitative research to explore and
hypermedia development practice of Croatian web de- analyze communications issues in both web team
signers/developers and are in line with the findings of and customer-to-designer relations;
other studies [1, 11, 21]. A Croatian web team con- • providing revised and renewed curriculum to stu-
sists of 3 designers/developers who develop Web or dents and workshops to practitioners so as to cover
multimedia systems through a process with clearly de- different issues in web/hypermedia development:
fined phases and activities, using mainly tool-driven specialized development methods, good design
development approaches, as well as informal and practice, usability and accessibility, web metrics,
semi-formal diagramming design techniques. The project management, communications skills etc.
survey analysis showed that the development process • introducing current and future clients with the
is not as undisciplined as it is generally considered. complexity of web/hypermedia development by
Moreover, it comprises development phases and ac- promoting scientific findings in IT magazines and
tivities identified in practice, similar to those in spe- newspapers, which could help in bridging the
cialized web/hypermedia design methods. However, communication gap among them and web profes-
the use of proprietary development methods is scarce sionals.
and the vast majority of designers/developers use in-
house or hybrid methods, which rely on tool-driven 6. References
development approaches.
A typical web project includes scope planning and [1] Barry, C. & Lang, M. (2001) A Survey of Multi-
requirements documentation, yet with low tendency to media and Web Development Techniques and
use written procedures or standards/guidelines/rules in Methodology Usage. IEEE Multimedia, 8(3), 52-
the process of web site development. The biggest 60.
problems in web or multimedia systems [2] Barry, C. & Lang, M. (2003) A Comparison of
development are communication with clients and con- "Traditional" and Multimedia Information Sys-
trolling the project time, cost and scope. tems Development Practices. Information &
Therefore, to deal with the recognized problems, Software Technology. 45(4), 217-227.
several areas of action which could help all stake- [3] Bichler, M. & Nusser, S. (1996) Modular Design
of Complex Web-Applications with W3DT.
 IEEE 5th Workshops on Enabling Technologies: World Wide Web Design. Retrieved May 12,
Infrastructure for Collaborative Enterprises, 2006, from http://citeseer.ist.psu.edu/
Stanford University, California, June, 19-21. kyaw98survey.html
[4] Ceri, S., Fraternali, P., Bongio, A. (2000) Web [17] Murugesan, S. & Ginige, A. (2005) Web Engi-
Modeling Language (WebML): A Modeling neering: Introduction and Perspectives. In Suh,
Language for Designing Web Sites. In Confer- W. (ed.) Web Engineering: Principles and Tech-
ence Proceedings of Ninth World Wide Web Con- niques, 1-30. Idea Group Publishing.
ference (WWW9), Amsterdam, May 15-19, 2000 [18] Nanard, J., Nanard, M. (1995) Hypertext design
[5] De Troyer, O. & Leune, C.J. (1998) WSDM: A environments and the hypertext design process,
user centered design method for Web sites. Com- Communications of the ACM, 38 (8), 49–56.
puter Networks and ISDN systems, Proceedings [19] Olsina, L.(1998) Building a Web-based informa-
of the 7th International World Wide Web Confer- tion system applying the hypermedia flexible
ence, Elsevier, 85-94. process modeling strategy. In Proceedings of
[6] Garzotto, F., Paolini, P. & Schwabe, D.(1993) Workshop on Hypermedia Development Proc-
HDM: A model-based approach to Hypertext ap- esses, Methods and Models (at ACM Hypertext
plication design. ACM Transactions of Informa- '98), Pittsburgh, Pennsylvania, USA, June. ACM
tion Systems, 11(1), 1-26. Press.
[7] Ginige, A. & Murugesan, S. (2001) Web Engi- [20] Powell, T.A., Jones, D.L. & Cutts, D.C. (1998)
neering: An Introduction. IEEE Multimedia. Web Site Engineering: Beyond Web Page Design.
January-March, 14-18. Upper Saddle River, NJ: Prentice Hall
[8] Isakowitz, T., Stohr, E.A. & Balasubramanian, P. [21] Rosson, M.B., Ballin, J., Rode, J. & Toward, B.
(1995) RMM: A Methodology for Structured (2005) ´Designing for the Web´ revisited: A sur-
Hypermedia Design. Communications of the vey of informal and experienced web developers.
ACM, 38(8). In Lowe, D. & Gaedke, M. (eds.): International
[9] Lang, M. (2002) A Survey of Hypermedia Sys- Conference on Web Engineering 2005. - Lecture
tems Design in Ireland, Department of Accoun- Notes in Computer Science 3579, Berlin:
tancy & Finance, National University of Ireland, Springer-Verlag, 522-532.
Galway. Retrieved June 16, 2006, from [22] Russo, N. L. & Graham, B. R. (1999) A First
http://www.is.nuigalway.ie/mlang/research/HM- Step in Developing a Web Application Design
QUESTIONNAIRE.pdf Methodology: Understanding the Environment.
[10] Lang, M. (2003) Summary Report on Hyperme- In Wood-Harper, A. T. et al., (eds), Methodolo-
dia Systems Design in Ireland. Department of gies for Developing and Managing Emerging
Accountancy & Finance, National University of Technology Based Information Systems: 6th In-
Ireland, Galway. July. Retrieved June 16, 2006, ternational BCS Information Systems Methodolo-
from http://www.is.nuigalway.ie/ gies Conference. pp. 24–33. London: Springer.
mlang/research/HM-REPORT-2003.pdf. [23] Schwabe, D. & Rossi, G. (1995) The Object Ori-
[11] Lang, M. & Fitzgerald, B. (2005) Hypermedia ented Hypermedia Design Model. Communica-
Systems Development Practice: A Survey. IEEE tion of the ACM, 38(8) August, 45-46.
Software, 20(2) March-April, 68-75. [24] Taylor, M. J., McWilliam, J., Forsythe, H. &
[12] Lang, M. & Fitzgerald, B. (2006) New Branches, Wade, S. (2002) Methodologies and Website De-
Old Roots: A Study of Methods and Techniques velopment: A Survey of Practice. Information
in web/hypermedia Systems Design. Information and Software Technology, 44(6), 381-391.
Systems Management Journal. 23(3) Summer, [25] Venable, J.R. & Lim, F.C.B. (2001) Develop-
62-74. ment Activities and Methodology Usage by Aus-
[13] Lowe, D. & Hall, W. (1999) Hypermedia & the tralian Web Site Consultants. 4th Western
Web: an engineering approach, John Wiley & Australian Workshop on Information Systems Re-
Sons, England. search, WAWISR 2001
[14] Koch, N. (1999) A Comparative Study of Meth- [26] Vora, P. (1998) Designing for the Web: A Sur-
ods for Hypermedia Development. Technical Re- vey. ACM interactions, 5(3) May-June, 13-30.
port 9905. Ludwig-Maximilians-Universität [27] Zhou, J. & Stålhane, T. (2004) Web-Based Sys-
München. tem Development: Status in the Norwegian IT
[15] Koch, N. & Kraus, A. (2003) Towards a Com- Organizations. In Bomarius, F. & Iida, H. (eds),
mon Metamodel for the Development of Web Product Focused Software Process Improvement:
Applications. In Cueva Lovelle, J.M., Gonzalez 5th International Conference, Profes 2004, Kan-
Rodriguez, B.M., Joyanes Aguilar, L., Labra sai Science City, Japan, April 5-8, 2004, Pro-
Gayo, J.E. & Paule Ruiz, M.P. (eds.), Third In- ceedings. Lecture Notes in Computer Science,
ternational Conference on Web Engineering Vol. 3009, 363-377. Heidelberg: Springer-
(ICWE 2003). Verlag.
[16] Kyaw, P. & Boldyreff, C. (1998) A Survey of Hy-
permedia Design Methods in the Context of