Introduction to Hypermedia

The term hypermedia comprises a set of ideas, although it seems to be understood somewhat differently within different disciplines. We will therefore start by examining what some of the more well-known researchers and developers of hypermedia say about it, in order to find some typical, common features.

The very first person to directly formulate the hypermedia concept was Vannevar Bush. President Roosevelt appointed him manager of the organisation which co-ordinated technological research in the USA during World War II. In 1945 he published the article "As we may think" in the magazine The Atlantic Monthly, addressing some of the problems faced by modern science. Even at that time the number of scientific publications was high, and Bush emphasised two resultant problems: finding the time to keep up-to-date professionally, and remembering all this information. He also pointed out the dynamic aspects of scientific publications, claiming (Bush p. 45):

"A record, if it is to be useful to science, must be continuously extended, it must be stored, and above all it must be consulted."

In proposing a solution he launched the idea of an imaginary machine to handle these problems, The Memex (Bush 45). The Memex was intended to structure information by association, in contrast to traditional structuring by index. Using links ("trails"), information units would be chained, and these were to be followed through a series of associations by pressing a button. The Memex was never realised, but many people consider this the precursor to hypermedia.

The concept of hypertext, and later hypermedia was created by Theodor Holm Nelson, the man behind the hypermedia system Xanadu. Fascinated by the ideas of Bush, from the 1960s onwards Nelson developed them further, placing the following content in the concept of hypermedia (Nelson 90, p. 0/2):

"Well, by 'hypertext' I mean non-sequential writing -- text that branches and allows choices to the reader, best read at an interactive screen."

As we see, Nelson assumes that both the reading and the writing process would be supported by hypertext. Nelson placed special emphasis on the non-sequential nature of hypertext. Information must follow its natural structure, not being constrained by the linear structure to which traditional writing must adhere, for example a textbook which has to be read from cover to cover. Generally this is also what Bush meant by following information via links ("trails"), while Nelson introduced the term links for these paths through information.

It is well worth noting that Nelson does not assume the use of information technology to represent hypermedia. Hypermedia is rather a method for structuring information. New information is accessed by "jumping" from a reference point directly to the referenced information. Nelson cites a newspaper front page as an instance of "manual hypertext" (Nelson 90, p. 1/17), the headlines with their page references being the reference points for the information found in the paper.

In parallel with Nelson, Douglas Engelbart researched some of the same ideas at the Stanford Research Institute (Engelbart 63) and (Engelbart 78). This yielded the development of NLS (oNLine System), later Augment. In contrast to Bush and Nelson, Engelbart assumed an overarching hierarchical structure of information in hypertexts, enabling links across these structures. In general, nonetheless, Engelbart's theories agree with those of Nelson and Bush.

Engelbart focused especially on the opportunities provided by hypertext for group work, and much of his research was on using hypertext as a tool for work groups. Several persons might share the same document, working on it at the same time. It would also be possible to work on the same document on the same screen at the same time, so that two persons would sit in different locations while seeing the same screen, which they could both edit. WYSIWIS -"What You See Is What I See". Furthermore, electronic mail was used to transfer documents and structures. All of these features have now been implemented in Augment (Engelbart 78).

Another major figure in hypermedia research, Jeff Conclin, offers a more pragmatic description of hypertext (Conclin 87, p. 17):

"The concept of hypertext is quite simple: Windows on the screen are associated with objects in a database, and links are provided between these objects, both graphically (as labelled tokens) and in the database (as pointers)."

This is a direct description of how a typical hypertext system organises information. Conclin assumes the use of a computer with a graphical interface. The information is stored in a database, suggesting that some typical database routines, such as taking back-up copies, data integrity and real-time transactions are handled for the user by the system.

Hypertext was originally constructed for purely text-based systems, but as computer features evolved, making it possible to represent other information as well, for example computer graphics, the term hypermedia was introduced. This concept was also created by Theodor Holm Nelson. As today's work stations are able to represent different types of multimedia, this difference has become less important. The rest of this text will thus employ the term hypermedia.

Based on the work of Bush, Nelson, Engelbart and Conclin we may list typical characteristics of hypermedia. These qualities largely agree with those of existing hypermedia systems:

The literature on this occasionally appears to use the terms hypertext and hypermedia in a wider sense, so that the requirements above are not met. We will use "hypermedia" in the stricter sense of the term in the rest of this lesson. Hypermedia is primarily a means for information exchange among persons.

At present work is underway to develop a common reference work for hypermedia, "The Dexter Hypertext Reference Model" (Halasz, Schwartz 94), which aims to standardise hypertext terms and concepts. A common standard may clarify terminology, permitting various hypersystems to be considered more homogeneously. Furthermore, such a standardisation may form the basis for exchanging hyperdocuments among different hypersystems. However, it is so far uncertain whether this standard would be appropriate for describing existing systems, as it may set too strict requirements, thus excluding some existing systems (Halasz, Schwartz 94, p. 38). Because of the lack of standards, formal descriptions and methods, hypermedia must be regarded as a still immature science. This would include both hypermedia itself as a technique and its applications. An important area for research is developing methods or models for hypermedia construction, and there is much work still to be done here. Initial work in this area suggests dividing these models into:

From the first category three models are especially familiar: The Hypermedia Design Model - HDM, The Hypermedia Data Model - HM-data model, and The Relationship Management Design Model - RMD. Examining these more closely would be beyond the scope of this course. The Department of Information Technology at NTNU is continuing its work on the further development of models for hypermedia systems. One post-graduate study is aiming to develop an object-oriented framework for hypermedia systems. Another is attempting to exploit rapid computer systems to navigate in normal data files as if they had been stored as hyperstructures. Data structure would then be generated on the fly based on a descriptive model.