Hypertext can be studied from many perspectives. First, there is the data modeling problem, regarding how to describe hyperbases (i.e., hypertext databases), and how to interchange them. All agree on nodes, anchors, links, paths / trails, webs / subgraphs but not on the details of their description nor on how to deal with composites. All agree that there is coupling between nodes and windows. All agree that selecting an anchor should lead to (possible choice among several links and) link traversal, with some related action resulting. However, the actions may vary depending on use of cards or scrolling, on type of media being presented, on whether the link is ``hard'' or ``virtual'', etc. No agreement has been reached regarding scripting languages, though the new ScriptX system from Kaleida Labs (supported by Apple and IBM) may emerge as a de facto standard.
Hypertext standards efforts do exist, both informally among leading researchers (working on the ``Dexter'' model) and formally (ISO has approved HyTime [5], for hypermedia and time-based performances), but they are either incomplete or very abstract, and the only agreement that may exist is regarding interchange. A recent workshop on Hyperbases identified numerous research issues regarding hypertext management systems, including version control, transaction management, support for IS&R, and support for collaboration.
A second key area is that of interfaces. This connects in with work on multimedia in terms of hypermedia systems. But even with hypertext, there are problems relating to navigation (see the Nielsen article for this unit), and of course, standards.
A third area is that of building hyperbases. To do this on a large scale requires automatic schemes to build links: that reflect the structure of documents (e.g., show inclusion relationships as from an SGML document), that handle references (e.g., bibliographic citations), that identify words to go into a glossary (e.g., by looking for a tag that indicated emphasis, or by noting phrases like ``is defined as''), etc. Content-based clustering should also be considered to help in the construction of maps, thesauri, and other supplements to hyperbases.
A fourth area is that of connecting to other software. There are tie-ins to database management systems to support the hyperbases. There are tie-ins to the operating system to support link services. There are tie-ins to the windowing system to support presentation and anchor location and processing.
A fifth area is that of applications. One of particular importance is to support collaboration, especially writing / argumentation / brainstorming. Another is for software libraries, to encourage reuse. Then there is our course theme, digital libraries!
In addition to the above, many other perspectives apply to this area! Hopefully this unit will introduce you to key issues and encourage you to use hypertext in the future, or to build improved hyperbase systems.