Peter Brusilovsky
Ian Beaumont

Hypertext/hypermedia systems and user-model-based adaptive systems (i.e. intelligent tutoring systems, information retrieval systems) are most often considered as being two different approaches to browsing information spaces and interface organisation. Adaptive hypertext and hypermedia systems (AHSs) attempt to bridge the gap between these approaches. AHSs enhance classic hypermedia with an intelligent agent which supports a user in her work with hypermedia. The intelligent agent can adapt the content of a hypermedia page to the user's knowledge and goals or suggest the most relevant links to follow. AHSs avoid the 'unrelevant adaptation' problem of classic adaptive systems by providing space for user-driven adaptation. And AHSs avoid the 'lost in hyperspace' problem of classic hypermedia systems by providing intelligent guidance.

The goal of the workshop was to provide a forum for the comparison of research experience in the development of adaptive hypermedia systems and for the discussion of several important topics centered around adaptive hypermedia. Foremost amongst these topics were:

1. Do we really need adaptive hypermedia systems?
Opponents hold that the very idea of hypermedia is that different users can adapt the hypermedia based system to their own needs, while the system itself should remain static.
2. What can be adapted in adative hypermedia.
Several ways of adaptive presentation support and adaptive navigation support were put up for discussion.
3. How the hypermedia can be adapted?
What types of user models and what adaptation techniques can be used.

The papers presented at the workshop and gathered in this volume range from position papers centering on the above topics to descriptions of implemented systems which have in some way confronted the adaptivity problem in hypertext systems. Brusilovsky's paper serves as an introduction and overview. Those of Beaumont and of Kay and Kummerfeld present teaching systems using hypertext. The domain of the former is brain anatomy, while that of the latter is programming in C, both systems being intended for university use. De Rosis, De Carolis and Pizzutilo present an approach for adaptive hypertext explanations in application systems. Zeilinger discusses the re-implementation of a standard CAL application (with user model) as a hypermedia system. Mayfield presents a system using a semantic net substructure for hypertext offering possibilities for adaptation to the user. The papers of Dufresne, Mathe and Chen, and Nill each describe a general approach to the problem.