Published on IEEE Computer, January 1997

Toward Support for Hypermedia on the World Wide Web

Michael Bieber
New Jersey Institute of Technology

Fabio Vitali
University of Bologna

0018-9162/97/$10.00 © 1997 IEEE
Copyright 1997 IEEE. Personal use of this material is permitted. However, permission to reprint/republish this material for advertising or promotional purposes or for creating new collective works for resale or redistribution to servers or lists, or to reuse any copyrighted component of this work in other works must be obtained from the IEEE.

As organizations rush to embrace the World Wide Web as their primary application infrastructure, they should not bypass the benefit of hypermedia support.

Many organizations will embrace the World Wide Web as their primary application infrastructure. However, in the rush to acquire and retrofit Web applications, they risk bypassing the Web's greatest supplemental benefit­hypermedia.

Hypermedia lets you add, access, and navigate links in textual and multimedia information. With hypermedia features, Web applications can provide clearer information, and they can be easier and more effective to use than traditional applications. We foresee a hypermedia-influenced era in which users will expect (and application developers will provide) browsing and supplemental linking in all applications.

However, some of the latest technologies may actually make it difficult to provide hypermedia features on the Web. For example, Java applets, which read and display some arcane binary data format as input parameters, will recreate the well-known problems with proprietary and unreadable data formats. Applets that create independent connections with a server application could provide shared-screen collaboration and other two-way communication, but they would require an independent and as-yet-undeveloped connection protocol that will not necessarily support hypermedia, rather than the standard HTTP.

Furthermore, even if designers agree that supporting a broad range of hypermedia features is desirable, they will need appropriate methodologies and tools. If designers must wait for these methodologies and tools, they may lose interest in using them and, therefore, in supporting hypermedia features. Thus, developing appropriate methodologies and tools quickly is imperative.


The World Wide Web's ubiquity and openness make it the choice for applications that require widespread distribution and access. In addition, extensive standardization and the fundamental modularity and simplicity of the Web's protocols and data formats make it the best environment for applications that are platform independent and/or that have been customized for international users. (See the sidebar, "The World Wide Web.")

With the next round of telecommunications infrastructure improvements, we assume that individuals and organizations will have Web access through their telephone or cable systems and that the Web will be the major application delivery and interface technology for years to come.

Web integration­in which users access an application, document, and so on through a standard Web browser­presents users, designers, MIS departments, and organizational management with a golden opportunity to enhance applications substantially with hypermedia features. This will become increasingly relevant as the Web matures and organizations feel secure in making it their primary application infrastructure.


While many users first experience hypermedia through the Web, hypermedia research has been ongoing since the early 1960s, when Doug Engelbart developed NLS, a multiuser distributed hypertext system.1 Since then, the hypermedia research community has developed many features, systems, guidelines, frameworks, and theories focusing on structuring, presenting, and accessing interrelated information.

Hypermedia is concerned with structuring and giving access to an application through its elements' interrelationships. Working with an application from this point of view can help developers understand it better and determine how they can make information technology more accessible and useful.

We take a broad view of relationships between application elements:

Intelligent hypermedia engines (which automatically create link anchors in documents and find link destinations for users) could use a design document that describes an application to infer all of these relationships among the applications' elements and processes, except ad hoc links declared by users.2 Figure 1 illustrates these relationships for a customer order invoice.

Figure 1. A typical product order invoice demonstrates the important relationships among application elements that are central to the use of hypermedia. The invoice is based on a standard template, so the user's system can identify what each piece of text represents. A hypermedia engine automatically infers all relationships, except ad hoc relationships. In this example, the schema relationship lets a billing analyst click on the customer's name and view additional information from the customer database. The information relationship lets a template developer see characteristics of the "name" field. The occurrence relationship lets an account representative see other templates with the customer's name, such as the accounting information template. The process relationship lets a systems analyst see other company operations that involve product quantities. The statistical relationship reveals other products often ordered with the product currently being requested, which lets a company representative ask if the customer also wants the other products. The structural relationship lets a user determine how the total price (amount) was calculated.

Relationship management

Hypermedia techniques provide the tools for relationship management,3 which entails creating, storing, maintaining, retrieving, filtering, customizing, presenting, commenting on, and navigating relationships.

A suite of hypermedia functionalities4 5 builds on relationships, augmenting information system applications with annotation, navigation, and structuring features. In structuring, you organize an application's components in terms of their interrelationships. For example, you provide links between documents, guided tours of related elements, and summaries of groups of related elements. Navigation allows various types of access­including information retrieval, browsing, and sophisticated backtracking­along each of these structural features.

Considering an application from a relationship management viewpoint results in new ways to

Design considerations. Web-based designers should at least consider including hypermedia features in every application. A system overview could present the application's major processes, and users could then zoom down to a particular task. There would be descriptions of all processes and commands. People could share annotations with teammates. In addition, users would have more control over their navigation through an application than they have now. Also, users could see hypermedia relationships among applications. This would help show how an object is used in a multiapplication analysis, or how information moves from one application to another and how each application then uses that information.

Collaborative hypermedia. To serve an organization's working groups and telecommuters, hypermedia features will have to coexist with collaborative features that let people work together on projects. Collaborative tools contain many relationships between a working group's members, tasks, and information. We thus view collaborative hypermedia,6 which promotes hypermedia support for collaboration as well as collaborative support for hypermedia, as an integral part of Web support.

Application use. Web environments should not overwhelm users by providing too many options. Web environments should include filtering mechanisms to present only the most relevant links, based on the users' current goals. Users could request other, out-of-the-ordinary information explicitly.

Too many options will slow applications that must function quickly (such as those used by telephone operators and order entry personnel) and make users less productive. Many applications, such as those for people who fill customer orders in warehouses, were designed to give users no options or decision-making power. Yet even in these jobs, anomalies occur. For example, a customer may have an unusual order that could be filled more quickly and cost-efficiently if warehouse workers had links to instructions for handling such a situation. Designers aware of hypermedia's potential could build these types of links into the system, thereby expanding user effectiveness and organizational efficiency.

Hypermedia implementation issues

The emerging field of organizational hypermedia7 8 considers how relationship management best serves information-processing within and among organizations. Someone in an organization must decide which applications will benefit enough from relationship management and other Web infrastructure features to implement in this environment.

Integration can be pushed by customers, system users, or the organization's MIS department. The decision will depend primarily on the ways integration will benefit individual applications and an organization's application suite, and on whether the appropriate development environment exists.

Designing robust hypermedia applications. While software engineering methodologies exist for standard application development, few design methodologies handle the linking and navigation inherent in hypermedia, and few specifically address Web standards or CGI (computer graphics interface) scripting. Current hypermedia design methods (discussed in the sidebar "Hypermedia design methodologies") produce only stand-alone, retrieval-oriented hypermedia applications.

Professional developers will perform most Web integration of major applications. However, end users­the majority of Web authors­currently perform most small-scale integration (such as converting documents and writing simple Java applets). Due to lack of training and the tedious nature of following formal methodologies, end users are no more likely to use a robust design methodology for integration than for developing a spreadsheet.

This often leads to poorly designed Web applications, poor performance and, to a lesser degree, poor appearance, all of which may discourage organizational acceptance of the Web. Developers should thus formulate, formalize, and develop a suite of analysis and design methodologies, CASE tools, and guidelines for the Web and relationship management to be used by full-fledged application developers and Web page authors.

Evaluating hypermedia applications. Robust applications require rigorous testing. Relationship management and other Web support will require a new suite of evaluation techniques to accomplish this. Few metrics and evaluation approaches exist for hypermedia-supported applications.

You can evaluate the actual document and link structure within an application. For example, you can determine whether all links have endpoints and whether users get to the information they want within a given number of link traversals.

You can also evaluate general accessibility. For example, you can determine whether users can easily access all related information for every object in a Web page, and whether users can clearly determine how the information is interrelated and in what context.

Rigorous application testing is critical to ensure performance in a networked environment, as well as link integrity (for example, whether destinations exist) and the presence of adequate contextual cues during Web page navigation. However, few people who aren't information systems experts know how to test applications rigorously.

Some researchers have started to work on hypermedia metrics,9 10 and some have conducted evaluation studies, although few actually use systematic evaluation methodologies. Companies that provide hypermedia conversion and organizational support have developed robust sets of automatic evaluation tools, but these are generally proprietary and unpublished.

Developers should continue trying to formulate and formalize evaluation approaches, and to produce automated evaluation software for Web-based applications.

Fourth-Generation Web-Support Features

Many consider the Web too complicated an environment within which to create or find information. In this regard, we have drawn an analogy between programming languages and hypermedia features.4 The current state of the Web is analogous to second-generation programming, in which programmers must specify all tasks in excruciating detail. Currently, developers and authors must build all hypermedia representations and navigation using single-step links without semantic or behavioral typing, and must construct any advanced features ad hoc in a programming language such as JavaScript.

We envision a fourth-generation Web environment that includes fourth-generation features and authoring tools. (Third-generation constructs specify program flow without specifying low-level details, while fourth-generation features and tools specify what they want to achieve instead of how the underlying program would produce it.) Developers are starting to design third- and fourth-generation tools now.

Providing high-level features and tools will boost the productivity of developers and users, making the Web as feasible a development environment as any that developers currently have off the Web. More fundamentally, however, programmers and authors will be able to take full advantage of the structuring, navigation, and collaborative support that hypermedia promotes.

Designers of future Web browsers and editors will build high-level features and tools into Web system software. Application builders, authors, and users will employ them. Application builders will incorporate some into their Web-based applications. Authors will use some to author document structures and collect reader preferences. Users will employ others to customize their interaction with Web applications.

Fourth-generation features

Fourth-generation hypermedia features would provide sophisticated relationship management and navigation support.

For example, the system could automatically display in a graphical overview the results of the hypertext structure query "show me all design documents with at least one negative annotation by the Legal or Public Relations Departments" divided into groups based on annotation author, document subject, or number of document links, as shown in Figure 2. Users could also open a document in the structural query results display and obtain a nested guided tour of the other documents that the query returned, as shown in Figure 3. This tour displays a copy of the document, shows the sections that drew comments, and provides a dialog box that lets users navigate easily through the structural query's results.

Figure 2. Structural query results. The user has asked for all documents with at least one negative annotation by the Legal or Public Relations Department. Documents have been shown in the "ranked results" display in order, based on how many qualifying negative annotations they had. In the "result cluster" display, employees of the Legal and Public Relations Departments are shown along with the number of documents (indicated by folded-page icons) they have commented on. Users can click on any icon to read the corresponding comment. In this figure, the document "Removing the Firewall" has been selected in the "ranked results" display. In the "result cluster" section, each employee's negative annotation on "Removing the Firewall" is shown in black. Users can read the annotation by clicking on the black icon. Meanwhile, users can further analyze "Removing the Firewall" in the "document link neighborhood" display, which shows each of the five negative annotations (indicated by typed-page icons) on the left and additional related information (three documents and a favorable annotation) on the right. Users can read any annotation or document by clicking on the appropriate icon.

Figure 3. Guided tour. Opening a document from the structural query results display in Figure 2 gives users a nested guided tour that displays a copy of the document, the sections that drew comments (each section is underlined and the type of comment appears in brackets), and a dialog box that lets users navigate easily through the various results of the structural query.

Systems could also provide templates of semantically typed nodes and links. For example, a Delphi-methodology-based system11 (which is designed to determine how a group of people feels about a particular issue) utilizing survey questions could dynamically generate the proper template for each question. Each question's template may have several interrelated parts, such as "arrange this list of factors in order of importance," "benefits of this factor," and "disadvantages of this factor." The system would generate structural links among a question's parts, schema links among questions, information links between an item and information about the item, and process links that determine the question that logically follows another question, based on previous answers.

Fourth-generation features include semantically typed documents and links, and links anchored in multimedia content.4

Fourth-generation tools

Fourth-generation hypermedia tools would help users manually and automatically generate fourth-generation hypermedia features. These tools should let users easily specify these features in a high-level, template-driven environment that permits programming by example or visual programming.

With these tools, users can specify parameters and behaviors for automatically generated features, such as the structural query and the guided tour shown in Figures 2 and 3, respectively. For manual construction, fourth-generation tools let users, for example, specify the components of guided tours and overviews. The fourth-generation environment must be simple enough to let any user construct these features.


MIS departments typically are responsible for engineering, development, and execution of software for large applications. They are also responsible for their organization's software and hardware environments, its computer-related training and help desks, and most other computing needs.

Because an organization's computer applications support its operations and fiscal prosperity, MIS managers must concern themselves with application stability, robustness, longevity, and security. MIS professionals test programs as exhaustively as possible to eliminate bugs that could slow or halt company operations. As people are usually the MIS department's most expensive resource, they need powerful development environments to maximize their productivity.

In its current state, the Web does not offer the highest levels of application stability, robustness, longevity, and security. Some organizations are making forays onto the Web, but most are not yet adopting it as their default programming environment.

MIS departments still lack

Developing these capabilities, developing effective standards, and solving the Web's well-known technical problems (including security, bandwidth, and scale) will contribute to the stability, robustness, longevity, and security of Web applications. MIS professionals will embrace the Web when this occurs and when they have appropriate design methodologies and productivity tools.

This will give MIS departments a common development environment, a common set of basic skills to expect from programmers, easy distribution of information, and sophisticated hypermedia support, in addition to the benefits of working on the Web.


While an organization's MIS department is concerned with application stability, upper management cares about the organization's prosperity and stability. Thus, many managers avoid the Web out of fear that they will lose control of the information that becomes public. Many companies thus use intranets, which provide organizational hypermedia features behind firewalls. Firewalls let organizations use Web technology and hypermedia creatively beyond public view.

Organizations are most concerned about operating on the Web because outsiders could annotate information that appears there. Third-party link services will allow consumer rating services, individuals, and competitors to annotate Web pages.12 Organizations whose pages have been annotated could respond, but the damage will have been done. Of course, consumer rating services, individuals, and competitors can express themselves about an organization now, but not from the organization's own Web site.

Organizations that move to the Web will find that hypermedia radically changes their application suite and slowly changes the way they do business. For example, with an easy distribution tool in place, companies may begin to engage customers more directly in marketing and decision making. A company could, for instance, use group coordination tools to offer a coupon to consumers who fill out a survey or participate in an online focus group. Organizations may also establish network tools that let consumers ask questions or make comments. In addition, by eliminating the need to talk to a customer service representative, more consumers might be encouraged to comment.

There are thus numerous advantages of moving to the Web. Organizations will have to decide whether they outweigh the disadvantages.

The World Wide Web infrastructure can serve as an interface to all interactive applications and, over time, will become the graphical user interface model for new applications.

Ubiquitous hypermedia support should become the jewel of the Web environment. Through Web integration, hypermedia could become an integral part of every interactive application. With the proper tools to support hypermedia in Web application development, it will become second nature for developers and individual authors to provide supplemental links and hypermedia navigation.

However, we fear that as organizations adopt the Web as their primary application infrastructure, designers will use Java and other tools to recreate current application functionality, and not take advantage of the Web's hypermedia-augmented infrastructure. If users, designers, MIS departments, and organizations don't demand hypermedia support, hypermedia may get lost in the frenzy of Web integration. v


We thank Mark Ackerman, Helen Ashman, John Avery, V. Balasubramanian, David Durand, David Mackler, Phil Murray, and Jim Whitescarver for their helpful comments. Our research was supported by the NASA JOVE faculty fellowship program, the New Jersey Center for Multimedia Research, the National Center for Transportation and Industrial Productivity, by New Jersey Institute of Technology Grant No. 991967, and by an AT&T Foundation grant.


1. D. Engelbart and W. English, "A Research Center for Augmenting Human Intellect," Proc. Fall Joint Computer Conf., 1968, pp. 395-410.
2. M. Bieber, "On Integrating Hypermedia into Decision-Support and Other Systems," Decision Support Systems, 1995, pp. 251-267.
3. T. Isakowitz, E. Stohr, and P. Balasubramanian, "RMM: A Methodology for Structuring Hypermedia Design," Comm. ACM, Aug. 1995, pp. 34-44.
4. M. Bieber et al., "Fourth Generation Hypermedia: Some Missing Links for the World Wide Web," Int'l J. Human Computer Studies, to appear.
5. J. Nielsen, Multimedia and Hypertext: The Internet and Beyond, AP Professional, San Diego, 1995.
6. A. Rana and M. Bieber, "Collaborative Hypermedia Education Framework," to be published in Proc. 30th Ann. Hawaii Int'l Conf. System Sciences, IEEE CS Press, Los Alamitos, Calif., 1997.
7. T. Isakowitz, "Hypermedia in Information Systems and Organizations: A Research Agenda," Proc. 26th Ann. Hawaii Int'l Conf. System Sciences, Vol. III, IEEE Computer Society Press, Los Alamitos, Calif., 1993, pp. 361-369.
8. M. Bieber and T. Isakowitz, "Introduction to the Special Issue: Hypermedia in Information Systems and Organizations," J. Organizational Computing and Electronic Commerce, Fall 1996, pp. iii-vii.
9. F. Garzotto, L. Mainetti, and P. Paolini, "Hypertext Design, Analysis, and Evaluation Issues," Comm. ACM, Aug. 1995, pp. 74-86.
10. E. Rivlin, R. Botafogo, and B. Shneiderman, "Navigating in Hyperspace: Designing a Structure-Based Toolbox," Comm. ACM, Feb. 1994, pp. 87-108.
11. M. Turoff and S.R. Hiltz, "Computer Based Delphi Processes," Gazing into the Oracle: The Delphi Method and its Application to Social Policy and Public Health, M. Adler and E. Ziglio, eds., Kingsley Publishers, London, 1995, pp. 56-88.
12. R.J. Schloss, "Novel Business Uses of Independently Created Hyperlinks in the World Wide Web: Basic Mechanism and Examples," Proc. 29th Ann. Hawaii Int'l Conf. System Sciences, Vol. II, IEEE CS Press, Los Alamitos, Calif., 1996, pp. 137-146.

Michael Bieber is an assistant professor of information systems in the New Jersey Institute of Technology's Computer and Information Science Department. At the school, he is also director of the Hypermedia Information Systems Research Lab and associate director of the Electronic Enterprise Engineering Program. His research interests include information systems, hypertext automation, multimedia, electronic enterprise engineering, and decision support.

Bieber has published numerous articles about hypermedia. He received a PhD in decision sciences from the University of Pennsylvania. He is a member of the ACM, the Association of Information Systems, the Society for Design and Process Science, and the Society of Information Management.

Fabio Vitali is a postdoctoral student at the University of Bologna, where he received an undergraduate degree in mathematics and a PhD in computer and law. His research interests include distributed hypertext systems, versioning systems, coordination technologies, and human-computer interaction. Vitali has published several articles and two books on distributed hypertext systems and versioning. He is a member of the IEEE and the ACM.

Contact Bieber at and Vitali at

First Sidebar

The World Wide Web

The World Wide Web was born as a medium for delivering and displaying hypermedia documents over a large network. It was created to let users access and display textual and, to a limited degree, image documents, not to let users generate distributed applications. However, several technology advances (from HTML forms and CGI to Java applets and ActiveX objects) now let developers create sophisticated applications.

Building blocks

The Web is the result of a series of applications, communication protocols, and data formats conforming to one or more sets of standards. These building blocks exhibit a number of peculiarities that directly affect the porting of applications to the Web.

Clients are applications that request resources, while servers are applications that provide resources. Some applications, such as proxies, act as both clients and servers, requesting resources on behalf of another application, and then delivering it to the original requester.

Resources can be documents, files, forms, query results, and so on. URIs (universal resource identifiers) serve as the names for such resources.

Many of these resources are formatted in HTML, a data format that lets documents contain raw text content, formatting instructions, and, most importantly, hypermedia link specifications (outgoing link parameters and destination URLs [uniform resource locators], and incoming link anchor labels). HTML also lets users place external objects (usually images) within documents.

The main purpose of the protocols supported by the Web is to deliver resources to requesters. Web clients can request or create hypermedia links to resources accessible through these protocols, such as FTP, NNTP (Network News Transfer Protocol), and Telnet, which were introduced in other environments.

The Web HTTP connection protocol is for use only in the Web environment. It is a MIME-compliant protocol that is

Web clients and servers use HTTP to communicate. HTTP proposes a very specific type of connection and cannot accommodate

Web limitations

Several limitations still prevent the implementation of some classes of applications. Some of these limitations are due to the current state of Web technologies, while others are inherent to the Web and will not be solved in the near future.

For example, HTML (Hypertext Markup Language) was developed to display and, to a limited degree, semantically structure text based on its role within the document, rather than its position in the final display. Therefore, HTML will not be able to produce highly graphical interfaces, precisely position interface objects, or support uncommon notations (such as those used in music or chemistry) for a long time.

The Web's problems are well known and solutions exist for several of them. For example, some browsers already support data formats that overcome HTML's limitations. In addition, Java is becoming an important tool for creating complex interfaces and using specialized communication protocols. It is vital, however, that these solutions maintain the ability to specify links, as HTML does.

Second Sidebar

Hypermedia Design Methodologies

Recently, several hypermedia researchers have developed formal design methodologies for hypermedia applications.1 These include the Enhanced Object-Relationship Model (EORM),2 the Hypermedia Design Methodology (HDM),3 the Object-Oriented HDM (OOHDM),4 the Relationship Management Methodology (RMM),5 and HyperDesign.6

Danny Lange of IBM has built software on top of the ONTOS object-oriented database system for EORM. EORM lets designers specify hypermedia links of any type and define hypermedia functionality within any standard, nonhypermedia object-oriented design methodology.2

HDM was the first major hypermedia design methodology to tailor database design techniques for hypermedia documents and links. Designers can use HDM to develop sophisticated navigation features. OOHDM extends HDM for use in object-oriented applications.

The most advanced design methodology, RMM, has an associated development package called RMCASE, which can generate a display-oriented stand-alone hypermedia application.7 The designer must model an application in terms of entities and relationships. RMCASE then generates an application with links among the related entities and among the attributes within each entity. RMM, which adds a user interface design focus to HDM, is the first complete design methodology. It covers the process from analysis to construction and testing.

HyperDesign provides a formal methodology that helps with user interface design for hypermedia applications. Developers can use the HyperDesign tool to identify 14 facets of user interface design, from determining an application's requirements to finalizing the screen layout. Previously, there was no systematic support for these creative aspects of the process.


1.Proc. Int'l Workshop Hypermedia Design 95, LIRMM, Montpellier, France, 1995.
2.D. Lange, "Object-Oriented Hypermodeling of Hypertext-Supported Information Systems," J. Organizational Computing and Electronic Commerce, Fall 1996, pp. 269-293.
3. F. Garzotto, L. Mainetti, and P. Paolini, "Navigation Patterns in Hypermedia Databases," J. Organizational Computing and Electronic Commerce, Fall 1996, pp. 211-237.
4. D. Schwabe, G. Rossi, and S.D.J. Barbosa, "Systematic Hypermedia Application Design with OOHDM," Proc. Hypertext 96, ACM Press, New York, pp. 116-128.
5. T. Isakowitz, E. Stohr, and P. Balasubramanian, "RMM: A Methodology for Structuring Hypermedia Design," Comm. ACM, Aug. 1995, pp. 34-44.
6. V. Balasubramanian and M. Turoff, "A Systematic Approach to User Interface Design for Hypertext Systems," Proc. 28th Hawaii Int'l Conf. System Sciences, Vol. III, IEEE CS Press, Los Alamitos, Calif., 1995, pp. 241-250.
7. A. Diaz et al., "RMC: A Tool to Design Web Applications," Proc. Fourth Int'l World Wide Web Conf., in World Wide Web J., Dec. 1995, pp. 559-566.

Back to the cover feature