The idea of hypertext and multimedia documents is not new; indeed, it has been discussed for decades. Two things occurred simultaneously in the early 1990s, however, that made these ideas much more interesting: the introduction of the World-Wide Web and its underlying HTML language, and the development of the Mosaic web browser for interpreting that language. These sparked a fire that became the internet revolution we see occurring today. The notion that resources should be networked, so that when a new one is created it becomes available to everyone immediately, and so that one always goes to the original source rather than a possibly outdated copy, may seem obvious to us now, but these were new ideas a decade ago. The fact that there was a free and universally available browser (originally Mosaic, but now Netscape and Internet Explorer) with a standardized language (HTML) meant that information exchange could occur without the usual platform-dependence issues associated with computer files. These are what made the web the success that it is today.
In 1993, the Geometry Center set up one of the first web sites in the United States, and I began to write hypertext documents. At that time, I made two predictions. The first was that a new way of organizing large amounts of data would need to be found (this was when Yahoo was just a small listing, and search engines didn't yet exist). The need for this today is even more clear, as search engines return thousands of matches for most queries, and categorized listings grow unmanageably large. Anyone who has ever tried to locate something on the web will agree that, while search engines and lists of links are good temporary measures, something better is needed for the future.
My second prediction was that the ability to write hypertext documents would lead to a new non-linear mode of writing that is significantly different from our current printed forms. This has not yet occurred on a large scale. While there are now many electronic mathematics journals, most of them are designed simply as a means of getting paper journals to their readers faster and cheaper. They do not take advantage of the capabilities of hypertext documents except in the most elimentary ways, and have not promoted a new idiom in writing. While these journals certainly are important, they still leave a new field of communication yet to be explored.
In 1994, I took my first step into that field by writing a research paper in a hypertext
format . Most peoples' idea of writing hypertext documents, especially then, is to write a linear one and then break it up into pieces that are linked together in a chain. I did not take that approach, but tried to see what this new technology had to offer. Most math papers begin with a series of definitions followed by preliminary results, and only at the end provide the main (and interesting) result of the paper; this is out of necessity, as one can not talk about an object intelligently until it is defined and its basic properties are determined. I took a different approach, however; I wrote about the interesting result first as though I were speaking to someone who understood the topic, but provided links to supporting material for those who might need additional information. This "layered" approach is, in my opinion, the most significant advantage that hypertext has to offer: you can write at a level of detail that makes sense for the topic, but provide additional detail for those who require it.
Some people seem to think that writing hypertext means you can do whatever you want, however you want, and that no organization is required. This is not the case. If anything, more organization is required. Since there are now many possible arrangements of the information, the choice of which one to use is crucial. The structure of a document itself carries information, so great care must be taken when developing the structure. It is certainly easy to write bad hypertext documents (just as it is easy to write bad linear ones); the goal is to develop methods of writing good ones. Linear papers have been around for thousands of years, and we have pretty well determined what works and what doesn't in this format; but we are only at the beginning of our understanding of hypertext writing styles, and it will be years before we have even the basics worked out (the prevalence of shoddy work on the web attests to this). If this medium is to become a useful one, we need to develop those styles and idioms. Perhaps we will find that the linear approach is the best one after all; but we should conclude that from a position of knowledge rather than ignorance.
Electronic Communications Projects
As scholars and teachers, we are presented with a formidable challenge in determining the appropirate and effective uses of electronic media in research and education. It is tempting to stick to the old, comfortable approaches that we are used to and ignore the possibilities being presented to us. A better choice, though, is to take up the challenge, and play an active role in developing these new idioms ourselves.