Concept webs are hypermedia documents containing definitions and discussions of the key concepts of a subject, internally linked according to conceptual structure and also linked to examples and to other related material, including such things as animations, simulations, experiments, interactive database searches, etc. The potential uses of concept webs are many, in education, research, and even commercial enterprise. The World Wide Web currently provides an excellent medium for dissemination of concept webs; cooperative development of concept webs is also supported reasonably well. This document briefly elaborates on the "concept of concept webs" and discusses their possible use as the first and most ubiquitous vehicle for "informating" of academic disciplines, by which we mean the development of both broad and deep computer representations of knowledge in the discipline.1 In particular, concept webs can play an important role in implementing an informatics curriculum at Rensselaer.
The development of concept webs can be structured as faculty/student projects. In this kind of project the technology is already available, though rapidly changing. What is needed is the considerable effort it takes to create, refine, and extend the concept web pages' content, including coping with the frequent changes in the underlying technology. Faculty alone cannot supply this effort, but students working with faculty guidance can. The projects could be carried out either as assignments in courses or as Undergraduate Research Projects or independent study projects.
In this early phase, definitions and discussion of concepts could be entered in natural language text without formal structure except for following some recommended guidelines for coverage, style, and types of links to related concepts and material. The primary goal would be human readability, with computer automated or semi-automated processing of the embodied knowledge a secondary goal to be achieved over a longer time frame.
One large existing web site that shows a lot of the potential of concept webs (though its author doesn't use that term) is SGI's Standard Template Library (STL) documentation, which makes extensive use of concept definitions and refinement in documenting the requirements and properties of a set of standard algorithms and data structures. See
where I've added a graphical representation of a major part of SGI's pages, with links into the pages.2http://www.cs.rpi.edu/~musser/concept_web,
The STL documentation example is an excellent but imperfect example of concept webs as I currently see them being developed as a means of informating academic disciplines. Indeed, we should probably recognize that concept webs will never be perfect; they can always be revised and extended, but in the educational context this is a virtue, not a drawback. For example, the STL web pages are static, but there are lots of possibilities for making them active (and interactive). The web pages for particular algorithms in the STL web pages could for example contain animations of the algorithms. They don't currently, but one can see good examples of such animations in the separate SGI Java Algorithm Demo site. Developing such active content is an area where students could be heavily involved.
This example is heavily compputer science-oriented, but the idea of hypermedia concept descriptions and refinement hierarchies can be applied to any subject area. One example I'd like to develop as a concept web is (an expanded version of) the definitions and discussion that follows.