Other Free Encyclopedias » Online Encyclopedia » Encyclopedia - Featured Articles » Contributed Topics from A-E

Designing Web-Based Hypermedia Systems - INTRODUCTION, CRITICAL REVIEW OF PRINCIPAL DESIGN ISSUES AND CHALLENGES, Visualizing the Structure of Hypermedia Systems

development techniques traditional methods

Michael Lang
National University of Ireland, Galway, Ireland


Although its conceptual origins can be traced back a few decades (Bush, 1945), it is only recently that hypermedia has become popularized, principally through its ubiquitous incarnation as the World Wide Web (WWW). In its earlier forms, the Web could only properly be regarded a primitive, constrained hypermedia implementation (Bieber & Vitali, 1997). Through the emergence in recent years of standards such as eXtensible Markup Language (XML), XLink, Document Object Model (DOM), Synchronized Multimedia Integration Language (SMIL) and WebDAV, as well as additional functionality provided by the Common Gateway Interface (CGI), Java, plug-ins and middleware applications, the Web is now moving closer to an idealized hypermedia environment. Of course, not all hypermedia systems are Web based, nor can all Web-based systems be classified as hypermedia (see Figure 1). See the terms and definitions at the end of this article for clarification of intended meanings. The focus here shall be on hypermedia systems that are delivered and used via the platform of the WWW; that is, Web-based hypermedia systems.

There has been much speculation that the design of Web-based hypermedia systems poses new or unique challenges not traditionally encountered within conventional information systems (IS) design. This article critically examines a number of issues frequently argued as being different—cognitive challenges of designing non-linear navigation mechanisms, complexity of technical architecture, pressures of accelerated development in “Web-time” environment, problems with requirements definition, the suitability of traditional design methods and techniques, and difficulties arising out of the multidisciplinary nature of hypermedia design teams. It is demonstrated that few of these issues are indeed new or unique, and clear analogies can be drawn with the traditions of conventional IS design and other related disciplines.


Visualizing the Structure of Hypermedia Systems

Essentially, hypermedia attempts to emulate the intricate information access mechanisms of the human mind (Bush, 1945). Human memory operates by associating pieces of information with each other in complex, interwoven knowledge structures. Information is later recalled by traversing context-dependent associative trails. Hypermedia permits the partial mimicry of these processes by using hyperlinks to create non-linear structures whereby information can be associated and retrieved in different ways. Otherwise put, hypermedia facilitates multiple paths through a network of information where there may be many points of entry or exit. This especially is the case with Web-based hypermedia, where users can enter the system through a variety of side doors rather than through the front “home page”. The undisciplined use of hyperlinks can lead to chaotic “spaghetti code” structures (de Young, 1990). As systems scale up, this causes the substantial problem of “getting lost in cyberspace,” whereby it becomes very difficult to locate information or navigate through the labyrinth of intertwined paths (Otter & Johnson, 2000; Thelwall, 2000).

Two principal reasons explain why difficulties in visualizing the structure of a hypermedia system may arise. First, non-linear navigation mechanisms lead to intricate multi-dimensional information architectures that are hard to conceptualize. Second, Web-based hypermedia systems are typically an amalgam of many different interconnected components, such as static Hypertext Markup Language (HTML) pages, client-side applets or scripts (e.g., Java, Javascript), dynamically generated pages (e.g., PHP, Perl, Active Server Pages, ColdFusion), media objects (e.g., JPEG, VRML, Flash, Quicktime) and back-end databases. Flows and dependencies are not as visible in Web-based hypermedia systems as they are for most conventional systems, and it can be quite difficult to form a clear integrated picture of the technical architecture (Carstensen & Vogelsang, 2001).

However, the phenomenon of systems being constructed using a multiplicity of components is not unique to Web-based hypermedia. In conventional systems design, tiered architectures that separate data, logic and interface layers are commonly used to assist seamless integration. One such approach is the Model-View-Controller (MVC) framework, which has also been found beneficial in Web-based hypermedia design (Izquierdo, Juan, López, Devis, Cueva & Acebal, 2003). Nor is the difficulty of designing non-linear navigation mechanisms unique to hypermedia. Within traditional printed media, certain types of material are intentionally designed to be used in a random-access non-linear manner, such as encyclopediae, thesauruses and reference works. According to Whitley (1998), hypermedia systems are different from other types of software applications because “the developers have to set up a number of alternatives for readers to explore rather than a single stream of text” (p. 70). This may be a new concept in software design, but elsewhere, technical writers have long experience of setting up multiple navigable paths in the design of electronic documentation, such as online help systems. It has been found that technical writing techniques can readily be adapted to navigation design for Web-based hypermedia systems (Eriksen, 2000).

Accelerated Development Environment

The capacity of organizations to respond and adapt quickly to rapidly changing environments is a well-recognised strategic issue. Accordingly, IS need to be flexible and able to adapt to changing business needs. Looking at trends in IS development over the past 20 years, project delivery times have dramatically shortened. In the early 1980s, Jenkins, Naumann and Wetherbe (1984) reported that the average project lasted 10.5 months. By the mid-1990s, the duration of typical projects had fallen to less than six months (Fitzgerald, 1997), and average delivery times for Web-based systems are now less than three months (Barry & Lang, 2003; Russo & Graham, 1999). These accelerated development cycles have given rise to the notion of “Web time” or “Internet speed” (Baskerville, Ramesh, Pries-Heje & Slaughter, 2003; O’Connell, 2001; Thomas, 1998), a development environment that is supposedly characterized by “headlong desperation and virtually impossible deadlines” (Constantine & Lockwood, 2002, p. 42).

Such compressed timeframes are made possible by the combined effect of two factors. First, modern-age, rapid-application development tools greatly speed up the development process, although it is sometimes argued that What-You-See-Is-What-You-Get (WYSIWYG) visual design tools invite a reckless “just-do-it” approach without much, if any, forethought. Second, the Web is an immediate delivery medium which, unlike traditional IS and off-the- shelf software applications, is not impeded by production, distribution and installation delays. Web-based systems can be easily and quickly launched by developing functional front-end interfaces, powered by crude but effective back-end software, which can later be modified and enhanced in such a manner that end users may be oblivious to the whole process.

Again, however, this phenomenon of reduced cycle times is not specific to Web-based hypermedia design, for it also affects the design of conventional systems (Kurata, 2001). Yourdon (1997) defined “death march” projects as those for which the normal parameters of time and resources are reduced by a factor of one-half or more. Such scenarios are now common across software development in general, not just Web-based systems. This is reflected by the growing interest amongst the general community of software developers in high-speed approaches such as agile methods, Rapid Application Development (RAD), timeboxing and commercial off-the-shelf (COTS) application packages. Indeed, one could say that this trend towards shorter cycles is reflective of a greater urgency in business today, brought about by dramatic advances in technology and exemplified by practices such as just-in-time (JIT) and business process re-engineering (BPR). Rapid flexible product development is a prerogative of the modern age (Iansiti & MacCormack, 1997). Considered thus, the phenomenon of “Web time” is not unique to Web-based hypermedia design, and it ought be regarded as an inevitable reality arising out of the age-old commercial imperative to devise faster, more efficient ways of working.

Requirements Elicitation and Verification

Traditionally, IS have served internal functions within organizations (Grudin, 1991). In contrast, the Web has an external focus—it was designed as a public information system to support collaborative work amongst distributed teams (Berners-Lee, 1996). As traditional IS are ported to the Web, they are turning inside-out and taking on a new focus, where brand consciousness and user experience design become important issues. In a sense, Web-based systems are shop windows to the world. Russo and Graham (1999) make the point that Web applications differ from traditional information systems because the users of Web applications are likely to be outside of the organization, and typically cannot be identified or included in the development process.

It is plainly true that for most Web-based systems, with the obvious exception of intranets, end users are external to the organization. Collecting requirements from a virtual population is difficult, and the same requirements elicitation and verification techniques that have traditionally been used in software systems design cannot be easily applied, if at all (Lazar, Hanst, Buchwalter & Preece, 2000). Although this is new territory for IS developers, the notion of a virtual population is quite typical for mass-market off-the- shelf software production and new product development (Grudin, 1991). In such situations, the marketing department fulfils a vital role as the voice of the customer. For example, Tognazzini (1995) describes how a team of designers, engineers and human factors specialists used scenarios to define requirements based on an understanding of the profiles of target users as communicated by marketing staff. Thus, marketing research techniques can be used in conjunction with user-centred requirements definition techniques to understand the requirements of a virtual population. To verify requirements, because end users can’t readily be observed or interviewed, techniques such as Web log analysis and click tracking are useful (Lane & Koronois, 2001). The use of design patterns—tried and tested solutions to recurring design problems—is also advantageous (Lyardet, Rossi & Schwabe, 1999).

Applicability of Traditional Methods and Techniques

It is often argued that approaches and methods from traditional systems design are inappropriate for Web-based systems (Russo & Graham, 1999; Siau & Rossi, 2001). Murugesan, Deshpande, Hansen and Ginige (1999) speak of “a pressing need for disciplined approaches and new methods and tools,” taking into account “the unique features of the new medium” (p. 2). It is arguable whether many of the features of Web-based hypermedia are indeed unique. Merely because an application is based on new technologies, its design should not necessarily require an altogether new or different approach. It may well be true that traditional methods and techniques are ill-suited to hypermedia design. However, for the same reasons, those methods can be argued to be inappropriate for conventional systems design in the modern age (Fitzgerald, 2000). Modern approaches, methods and techniques—such as rapid prototyping, incremental development, agile methods, use cases, class diagrams, graphic user interface (GUI) schematics and interaction diagrams—are arguably just as applicable to hypermedia design as to conventional systems design. Methods and techniques from other relevant disciplines such as graphic design and media production also bear examination, as evidenced by the findings of Barry and Lang (2003).

Diagrammatic models are often useful in systems design to help overcome the cognitive difficulties of understanding complex, abstract structures. It has been argued that diagramming techniques from traditional systems design are inappropriate for modelling hypermedia systems (Russo & Graham, 1999; Siau & Rossi, 2001). One could just as easily argue that the flow of control in modern visual event-driven and object-oriented programming languages (e.g., Microsoft Visual Basic, Borland Delphi, Macromedia Lingo) is such that traditional techniques such as structured flowcharts and Jackson Structured Programming (JSP) are of limited use. For these types of applications, modern techniques such as Unified Modelling Language (UML) are being used, as well as approaches inherited from traditional dynamic media (e.g., storyboarding). Both storyboarding and UML can likewise be applied to hypermedia design; indeed, a number of UML variants have been proposed specifically for modelling hypermedia systems (Baumeister, Koch & Mandel, 1999; Conallen, 2000).

Multidisciplinary Design Teams

Perhaps the only aspect of Web-based hypermedia systems design that is radically different from conventional systems design is the composition of design teams. In conventional systems development, designers tend to be primarily “computer professionals.” This is not the case in hypermedia systems design, where team members come from a broad diversity of professional backgrounds, many of them non-technical. The challenge of managing communication and collaboration within multidisciplinary design teams is by no means trivial, and if mismanaged is potentially disastrous. Experiences reveal that discrepancies in the backgrounds of team members can give rise to significant communication and collaboration problems (Carstensen & Vogelsang, 2001). The multidisciplinary nature of design teams must be acknowledged in devising mechanisms to overcome the challenges of Web-based hypermedia design. Integrated working procedures, design approaches, diagramming techniques, toolset selection and mechanisms for specifying and managing requirements must all take this central aspect into consideration. The two foremost disciplines of Web-based hypermedia design are software engineering and graphic design (Lang, 2003), but alarmingly, it has been observed that these two factions have quite different value systems (Gallagher & Webb, 1997) and “appear to operate in distinctly different worlds” (Vertelney, Arent & Lieberman, 1990, p. 45). This is a considerable challenge which, if not addressed, could foil a project. Lessons can be learned from other disciplines that have successfully balanced the relationship between critical functionality and aesthetic attractiveness, such as architecture/civil engineering, automobile design and computer game development.


Throughout the history of computer systems design, it has been common amongst both researchers and practitioners to greet the arrival of much-hyped next-generation technologies by hailing them as profound advances that warrant entirely new approaches. Web/hypermedia design is another such example. However, Nielsen (1997) has commented that “software design is a complex craft and we sometimes arrogantly think that all its problems are new and unique” (p. 98). As this article reveals, few of the challenges of Web-based hypermedia design are indeed new or unique. Parallels can be drawn with lessons and experiences across a variety of disciplines, yet much of the literature on hypermedia design fails to appreciate the wealth of this legacy. Design methods, approaches and techniques can be inherited from many root disciplines, including traditional IS development, software engineering, human-computer interaction (HCI), graphic design, visual communications, marketing, technical writing, library information science, media production, architecture and industrial design. To paraphrase a well-known saying, those who choose not to draw from the well of cumulative knowledge are bound to foolishly repeat mistakes and to wastefully spend time seeking solutions where they might already exist. This article, therefore, concludes with a petition to hypermedia design researchers that they resist the temptation to dub themselves a “new” discipline (Murugesan et al., 1999), and instead reach out to explore the past and present experiences of related traditions.

Desktop Virtual Reality [next] [back] Desensitization and Media Effects

User Comments

Your email address will be altered so spam harvesting bots can't read it easily.
Hide my email completely instead?

Cancel or