Hypertext: A Critical Analysis
John C. Bradley
Hypertext is a computer-based system for organizing, storing, and accessing print-based information in a non-sequential manner. This paper discusses the hardware that makes hypertext possible and describes some prominent hypertext software. The psychological basis for hypertext are reviewed--schema theory, web learning, and generative learning. Issues related to hypertext as a learning medium are discussed, as are implications for instructional design.
Issue Addressed by Paper
Jonassen (1986, 1988, 1991) describes hypertext as a computer-based software system for organizing, storing, and accessing information in a non-sequential manner. Hypertext is dynamic rather than static. Hypertext allows for multiple pathways by readers with different interests, permitting them to determine their own individual presentation sequences based on their own preferred styles of reading or information needs. Hypertext gives the user immediate access to any of the information in a knowledge base.
Each chunk of information is called a node. These nodes are connected by "links," which are generally key words identified by highlighting, color coding, or reverse video. Sometimes links are identified as icons embedded in the text.
Jonassen (1986) describes three levels of hypertext: Chucked hypertext is completely random. Each node carries the same weight and can be accessed from any other node. Structured hypertext consists of sets of related nodes, each set accessible form any other set. In hierarchical hypertext, more detailed nodes are subsumed under more general concepts, requiring users to move up and down through the hierarchy in order to access related concepts.
Jonassen identifies several general applications for hypertext systems: Browsing systems enable users to navigate and explore information knowledge bases. Problem exploration systems are work-related and facilitate specific task domains. These systems help users organize and construct information. Macro-literary systems are not single documents but are a collection of material linked together by hypertext. General-purpose systems are not designed to facilitate any specific function, but can be tailored to a variety of needs.
Purpose of Paper
This paper will provide an overview of the technological, psychological, and pedagogical perspectives related to hypertext as a learning medium. Implications for instructional design are discussed. Significant issues are raised and conclusions are made.
Critical Review and Analysis
Although the concept of hypertext in print form was proposed forty years ago, the advent of microcomputers has made hypertext technologically feasible (Marchionini, 1988). Microcomputers have a variety of attribute that make them ideal for hypertext.
Computers can store large amounts of data in digital, electronic form. The compact disc with read-only memory (CD-ROM) is a well-established new technology. Each side of a 4.75-inch CD has the capacity to store 150,000 pages of text. (Locatis, 1989)
Computers can be networked into even larger databases. Commercial systems such as TELENET, and university systems like Internet, link computers and terminal together by telephone, allowing individuals to access different databases, send electronic mail, and browse bulletin boards. Although it has been difficult in the past for users to access multiple information systems--since the systems use different operating systems and commands--gateway software is being developed which will translate commands across systems. With such software, users will be able to log onto a single system to obtain information from different systems. (ibid.)
Computers are fast. Microcomputers with processing speeds of 10 megahertz are common. Even faster devices exist (ibid.). Such speeds make it possible to search massive databases for key words in a short period of time. This speed also allows for rapid, arbitrary jumps to material stored at another location in the computer's memory. (Jonassen, 1986)
High-resolution screens allow for the simultaneous display of text (in several sizes and fonts) and graphics (still and animated). Icons, color, and reverse video can be used to signal users about a hypertext link. Information displayed on a screen can be scrolled smoothly or flipped instantaneously. Windowing, or simultaneous display of more than one document on a screen is also possible. (Jonassen, 1986)
There are a number of software programs that exemplify the current capabilities of hypertext. KMS is a hypertext and hypermedia program created by Knowledge Systems, Inc. It operates on high-end workstations. A highly flexible tool for both writers and readers, KMS supports information creation, editing, and retrieval in networks. The system is explicitly aimed at improving productivity among groups that use very large databases or complex information (such as policy and maintenance manuals, project managers and expert systems). Versions of KMS have been used on the nuclear aircraft carrier, USS Carl Vinson, and in Westinghouse nuclear power plants. (Binder, 1989)
Another system designed for high-end workstation is Document Examiner, developed by Symbolics, Inc. It is used for presenting technical and operation manuals to users of Symbolics' artificial intelligence computers. It provides access to thousands of pages of technical information by combining both associative and search technologies with an easy-to-operate user interface. Users can search by keyword to locate likely entry-points into the documentation, then traverse through the database by using the mouse to follow hypertext linkages from that point on. Document Examiner was created specifically to provide efficient access to technical documentation. A measure of this system's success is that as many as 40 percent of the system's users do not even remove the shrink wrap from Symbolics' paper manuals because they are able to find what the need so readily using Document Examiner. (ibid.)
Window Book, from Box Company, was the first commercially available hypertext system for IBM-compatible microcomputers. The developer of Window Book, Michael Spier, aimed to produce an efficient tool for early IBMs--which had limited storage and graphics capabilities. Window Book was not developed specifically as a hypertext system, but as a cross-referencing system for large diskette-based documents. Eventually, Box Company was commissioned to expand the system's capabilities for navigating through hypertext documents published on multi-volume CD-ROMs. These document are often tens of thousands of pages in length. Among the system's strengths is a set of built-in navigational aids. This includes an automatically created table of contents, which is a "bread crumb trail" that records what articles the user has accessed in traveling through hyperspace. Licensees can choose to integrate Window Book in a context-sensitive manner with other software or hardware systems, to create links from inside a Window Book to bit-mapped graphics or interactive videodisc, or to provide full-text search or keyword search capabilities. There are versions of Window Book that run under MS-DOS, Microsoft Windows, and the Unix operating system. (ibid.)
Guide is a hypertext system released by Office Workstations, Ltd. It was originally developed for the Macintosh, but was later adapted for IBM PCs operating under Microsoft Windows. Guide was not originally developed as a hypertext system, but as an experiment in putting text into electronic form. In fact, Guide is much like a standard word processor. Because it operates in a bit-mapped graphics environment, Guide is capable of including complex diagrams and technical drawing along with textual materials. It employs a mouse-driven user interface with icons and pulldown menus. Ford Motor Company has used an industrial version of Guide for its database of technical manuals. Unfortunately, Guide suffers form a lack of efficient built-in navigation aids, a weakness that allows users to become easily lost. (ibid.)
Hypercard is perhaps the best known and most-used hypertext and hypermedia system. Hypercard is shipped free with each Apple Macintosh computer. Hypercard is both more and less than a true hypertext system. Because it includes a linking feature for creating paths through textural, graphic, and multimedia information, Hypercard can function as a text-retrieval tool. In fact, many of the commercial application of Hypercard have been textual publication, usually including substantial amounts of graphics as well. One example is the Hypercard Whole Earth Catalog. Hypercard includes basic, but slow, text search capabilities. Several third-party developers have offered faster add-ons for text search. Hypercard functions well as a front-end to complex networks of other computer programs, databases and system softwares. Hypercard has its own programming language (HyperTalk) that can be used to customize Hypercard for a wide-variety of uses. (ibid.)
One of the theoretical bases for hypertext is schema theory. A person's schema for an object, event, or idea consists of the amalgamation of distinctive features of and associations to the idea. Each schema represents a mini-framework on which to interrelate elements of information about a topic into one conceptual unit. Schemata are linked one to another by context-dependent descriptions. That is, the relationships between any two schemata are relative to the context in which they are used. One schema refers to another only through the use of a description which is dependent on the context of the original reference. In different contexts, people will use different schemata. The schemata that a learner accesses to interpret new information is necessarily idiosyncratic. Hypertext is a form of communication which accommodates these idiosyncrasies. (Jonassen, 1986)
A learner's s schematic network is often diagrammed and represented spatially as webs of information. Web learning principles assume that new information is integrated into prior knowledge by means of a web of structures rather than in a linear fashion. New material is intertwined in the web at nodes (schema) that are related to it. The web grows as learners acquire more detailed information. In order to connect new information to the learner's existing structure, we must first present a supporting web structure, then add details later. Hypertext manifests web teaching principles by using hierarchical connections between nodes. (ibid.)
Similar to web learning theory is the generative learning hypothesis. Generative learning principles contend that when interacting with information or instructional stimuli (text, illustrations, language), learners activate prior knowledge structures for the purpose of interpreting the stimuli. This is, new information has meaning only insofar as learners can find prior knowledge to explain it. According to the generative model of learning, learning is an active process of constructing knowledge. What an individual comprehends from material depends on what they already know. How they interpret information, then, depends on what they know, how it is organized, and how they are able to access it and relate it to new knowledge. For example, the meaning of a text is determined by the learner. Given the same textual material, each of us will generate somewhat different interpretations for it. Hypertext permits learners to individualize the knowledge-acquisition process. Hypertext allows users to interact with new information in the way that is most useful to them--that is, to customize the accession of information. (ibid.)
Wilson and Jonassen (1989) recommend that learners be allowed to make as many control decisions as possible in a hypertext environment. Learners should be provided with learning aids and expert advice only when necessary.
However, effective self-management skills are a factor in how much control the learner should be given in hypertext versus how much guidance the system should provide via tutorials and suggest paths (Kinzie & Berdel, 1990, and Wilson & Jonassen, 1989). Less structure should be provided for high-ability students and more for low-ability students (Allred & Locatis, 1988).
An important issue is integration of the hypertext information with knowledge structure of the learner. The less structured the hypertext is, the less likely users are to integrate what they have learned. Without an explicit external organization, many learners have difficulty acquiring new knowledge. (Jonassen, 1988)
Distraction may result from the high level of learner control that hypermedia systems provide. Cognitive resources may be diverted from content and relationships as learners attend to navigational decision-making. Distraction is also compounded by the vast quantities of information available at the click of a mouse. (Marchionini, 1988)
Related to distraction is the possibility of information overload. The richness of non-linear representations carries a risk for intellectual indigestion, loss of goal-directedness, and cognitive entropy. (Jonassen, 1988)
Disorientation is also a common problem for learners in a hypertext environment. Traveling from link to link to link may result in "getting lost in hyperspace. " When learners use a system, they need to know where they are in relation to some kind of identifiable structure. To move efficiently through hypertext and find needed information, learners must be able to return to a familiar origin or to see their location on some kind of map or table of contents. (Binder, 1989)
Implications for Instructional Systems
Jonassen (1986) has proposed a systematic procedure for the design of hypertexts: 1) Identify all key concepts. This is basically a content analysis. 2) Map the structure of the content. This includes tree construction, networking, and noting patterns. 3) Verify the structure by consulting subject-matter experts. 4) Determine the type of hypertext structure--chucked, structured, or hierarchical. 5) Prepare concept blocks. That is, write the text for each node. 6) Provide links and cues to other concepts. Links may be indicated by highlighted keywords or icons in the text. 7) Debug the system. Try out each and every option of the hypertext to be certain that the system performs faultlessly.
Binder (1989) offers a number of tips for hypertext writing. 1) Modularity. Information should be broken into small units. This eliminates the need for redundancy. Each module should be assigned a title or set of keywords. 2) Structured writing. A consistent style and format should be used for each chunk. 3) Linking. Links should be kept to a minimum. A link should only be made when there is a compelling reason to do so. 4) Layers of views. Since hypertext is modular, it is possible to create the effect of having different documents for different types of users. For instance, a personnel manual might be programmed to show different modules depending on whether the users was a low-level employee or a high-level supervisor.
On a more technological note, Binder (1989) recommends a number of navigational aids to help users maneuver through hypertext. 1) Graphical maps (a.k.a. graphical browsers) are a must. A good hypertext should begin by showing a map of the whole text. This map should represent the whole web of interrelated concepts contained in the hypertext. 2) A "bread crumb trail" is an aid that automatically records where the user has been in his or her movement through the database. For instance, Hypercard offers an iconic representation of the last 40 cards visited. A benefit of this feature is that a user may go back to past references without having to trace backwards through layers one-by-one. 3) Commands that enable users to step back or ahead one node at a time in a default sequence also prevents users disorientation. Hypercard uses arrow buttons as forward and back switches.
Binder (ibid.) also recommends link-level controls that allow a user to limit the links shown. By filtering the options down to a manageable size, the reader is better able to move through large hypertext documents.
Pulldown menus, such as those found in Macintosh and Microsoft Windows platforms, make accessing commands easy for users. Speed-keys are also recommended. (Wilson and Jonassen, 1989)
Some hypertext systems are designed to be freely interactive, allowing users to either read or write information at any time. In other types of applications, central control or security of the core document is an important consideration. Thus, a system providing multiple levels of access is often appropriate. (Binder, 1989)
In hypertext systems, a user should have quick access to on-line references. One way of providing this is context-sensitive help. That is, the system can provide information or appropriate assistance to the user depending one where the user is in the text or what commands the user is activating. This help may be requested by the user, or it may be unsolicited advice unknowingly activated by the users actions. (ibid.)
Kinzie & Berdel (1990) recommend other useful tools for hypertext users: glossaries of unfamiliar terms, note pads for users to clip and write textual information for later easy access, and drawing tablets for creating graphic notes.
Summary of Significant Issues and Conclusions
Technologically, hypertext is still in its infancy. As the capabilities of microcomputers increase, so will the sophistication of hypertext systems. Hypertext will eventually be part of a new generation of fully integrated electronic media. David Brin (1991) envisions a 21st century global hypermedia system that full integrates text, graphics, computer software, electronic mail, telephone, video, and audio. Instructional designers should keep abreast of technological progress but avoid becoming too much of an expert in any one medium.
Psychologically, we know that everyone learns in slightly different ways. Hypertext allows learners the freedom to learn in their own way. It enables the non-sequential exploration and assimilation of written information. However, since learners often require help in forming new knowledge structures, explicit organization and direction should be available when needed. The proper balance between discovery learning vs. directed learning in hypertext environments is one which deserves additional research by instructional designers.
Pedagogically, hypertext offers a number of challenges. How does one write objectives and create appropriate learning assignments for hypertext? How does one evaluate student learning in a hypertext environment? What is the proper role of the teacher? What is the proper role of the student? Here, too, instructional designers can help determine the effective application of hypertext in education.
- Allred, K.F. and Locatis, C. (1988). Research, instructional design, and new technology. Journal of Instructional Development, 11(1), pp. 2-5.
- Binder, C. (1989). Hypertext design issues. Performance Improvement Quarterly, 2(3), pp. 16-33.
- Jonassen, D.H. (1986). Hypertext principles for text and coursewear design. Educational Psychologist, 21(4), pp. 269-292.
- Jonassen, D.H. (1988). Designing structured hypertext and structuring access to hypertext. Educational Technology, November 1988, pp. 13-15.
- Jonassen, D.H. (1991). Hypertext as instructional design. Educational Technology, Research and Development, 39, pp. 83-92.
- Kinzie, M.B. and Berdel, R.L. (1990). Design and use of hypermedia systems. Educational Technology, Research and Development, 38, pp. 61-68.
- Locatis, C. (1989). Information retrieval systems and learning. Performance Improvement Quarterly, 2(3), pp. 4-15.
- Marchionini, G. (1988) Hypermedia and learning: freedom and chaos. Educational Technology, November 1988, pp. 8-12.
- Wilson, B.G., and Jonassen, D.H. (1989). Hypertext and instructional design: some preliminary guidelines. Performance Improvement Quarterly, 2(3), pp. 34-39.