Hypertext

Hypertext, made famous by the World Wide Web, is most simply a way of constructing documents that reference other documents. Within a hypertext document, a block of text can be tagged as a hypertext link pointing to another document. When viewed with a hypertext browser, the link can be activated to view the other document. Of course, if you're reading this document, you're already familiar with the concept.

Hypertext's original idea was to take advantage of electronic data processing to organize large quantities of information that would otherwise overwhelm a reader. Two hundred years ago, the printing press made possible a similar innovation - the encyclopedia. Hypertext's older cousin combined topical articles with an indexing system to afford the researcher one or perhaps two orders of magnitude increase in the volume of accessible information. Early experience with hypertext suggests that it may ultimately yield an additional order of magnitude increase, by making directly accessible information that would otherwise be relegated to a bibliography. Hypertext's limiting factor appears not to be the physical size of some books, but rather the ability of the reader to navigate increasingly complex search structures. Currently, additional increases in human information processing ability seem tied to developing more sophisticated automated search tools, though the present technology presents possibilities that remain far from fully explored.

Augmenting basic hypertext with graphics, more complex user input fields and dynamically generated documents adds considerable power and flexibility to this concept. Hypertext, though still useful for its original goal of organizing large quantities of information, becomes a simple, general purpose user interface that fits neatly into the increasingly popular client-server model. It does not seem difficult to image a day when restaurant orders, for example, will be taken using a hand-held hypertext terminal, relayed directly to the kitchen for preparation, and simultaneously logged to a database for later analysis by management.

Characteristics of good hypertext

The flexibility of hypertext gives free range to the author's creativity, but good hypertext appears to have some common characteristics:

• Lots of documents. Much of the hypertext's power comes from its ability to make large quantities of information accessible. If all the text in your system can be printed on ten pages, it would be just as simple to read through it from beginning to end and forget all this hypertext silliness. On the other hand, if there are ten million pages of text in your system, then someone could follows a link on atomic energy and ultimately hope to find a description of the U-238 decay process.

• Lots of links. If each document has just one link, then it is little more than normal, sequential text. A hypertext document should present the reader with several links, offering a choice about where to go next. Ideally, a document should present as many relevant links as the reader can easily comprehend and select among.

• Range of detail. The great advantage of hypertext is that it permits readers to explore to a breadth and depth that is simply not feasible in print. To make this accessible, available hypertext documents should range from the broadest possible overview of a subject, down to its gritty details. Imagine the Encyclopedia Britannica, all thirty-odd volumes of it, searchable online and with each article possessing links to a half dozen reference documents with even more detailed subject coverage. This is the potential of hypertext.

• Correct links. This may seem trivial, but it's amazing how many Web links point nowhere. In general, be careful linking to any hypertext document not under your direct control. Can you count on it to be there later?

Guidelines for a hypertext reference system

Hypertext seems best suited for reference material, so here are my suggested guidelines for creating hypertext reference systems, with the Internet Encyclopedia offered as an example:

• Reference documents. Start by assembling a good set of core reference material. In the Encyclopedia's case, the RFCs that document standard Internet protocols form this core. Ideally, the reference core should consist of extremely detailed documents, offering the highest possible degree of completeness. A general reference work on physics might start with a large collection of scientific papers.

• Topical articles. Augment the core reference material with articles at a broader level of detail. Systematic organization of these articles, perhaps using an outline as a framework, is essential to making them accessible to the reader. The articles should be focused, and short enough to be easily digestible in one piece.

• Search engine. A good search engine is invaluable for any large collection of documents. The Internet Encyclopedia uses a search model based on searching outward from a particular page, in order to facilitate both topical and keyword searches.

• Extras. These can include graphics, audio and video clips, problems and exercises, student courses, simulations, sample programs, ordering forms, database tables, and revision histories, to name a few.