From bob@cs.wisc.edu Tue Oct 4 19:00:32 1994 Date: Tue, 4 Oct 94 10:08:50 -0500 From: Bob Kummerfeld To: smart-ht@cs.umbc.edu Cc: bob@cs.wisc.edu Subject: submission for the Intelligent Hypertext workshop We wish to submit the following position paper to the Workshop on Intelligent Hypertext to be held in conjunction with CIKM'94. Postscript of the paper is available if you prefer. We also have several recent papers on our customised hypertext project if they are of interest - see http://emmetal.cs.wisc.edu:1994/bob.html Bob Kummerfeld (bob@cs.su.oz.au) Judy Kay (Judy@cs.su.oz.au) ----------------------------------------------------------------------- User models for customised hypertext J. Kay and R. J. Kummerfeld Department of Computer Science University of Sydney 1. Introduction An important use for hypertext is in the delivery of teaching material. Many computer aided learning systems use hypertext in one form or another allowing the user to pick their own path through the material, request further elaboration or descriptions of terms (glossaries). In general, however, there is only one description of each point and only one explanation of each term. If these don't suit the user, for reasons of background, preknowledge or learning style, then the hypertext will not be as effective as it could be. One solution to this problem is to provide multiple explanations and descriptions and allow the user to select the one that they find most appropriate. So, for example, a user who likes mathematical examples would choose one hypertext link, while a user who prefers text-based examples would select a different link. This means that the user that prefers mathematical examples will have to choose that option every time they meet a new concept. This approach may get very complex and unwieldy if many learning styles are accommodated. In the case of exercises designed to test understanding of the work it may be very difficult to guide the user to the one they are most suited to since it may depend on many factors in their pre- knowledge. Even simple variations in language level may become tedious for the user to select at each point. Our approach is to extend hypertext, customising it on the basis of a user model. We currently construct the model largely from a dialogue with the user before the begin the learning task. 2. User models to customise hypertext A user model is a collection of information about a user. In a hypertext for teaching, the user model would contain the user's preknowledge and preferences. As an example, the authors have constructed a hypertext course for the C programming language. The user model used in this system has information about what programming languages the user is familiar with and how well they know various concepts in those languages as well as the user's knowledge of relevant aspects of machine architecture. It also contains the user's preferences for presentation of information about programming languages: abstract versus concrete; terse versus more detailed; active learning versus directed. These preferences and other information about the user may come from the initial user model or may be inferred. For example, if a user says they are a competent Pascal programmer then it may be inferred that they understand concepts that are common to both Pascal and C. The tools for determining this and for storing the model are part of a toolkit for user modelling (Kay, 1990; 1994, Cook and Kay, 1994). Essentially, they use a range of sources to collect information. They also enable the user to inspect the user model and to contribute information to it. The information in the user model is used to customise the hypertext that is presented to the user. So, for example, a user who knows Pascal well, likes terse, abstract explanations that use jargon as appropriate will navigate a hypertext that is in this form. 3. Architecture In our system, customised hypertext pages are constructed by an intermediary program that is invoked when a user selects a link. This program consults the user model and, guided by what it finds, modifies the text of the target hypertext page to suit the user. In our first implementation we have used World Wide Web tools (server and browser) and the C preprocessor. Our customisation program uses a user modelling toolkit to take preferences from the user model and add them as "#define" statements to the raw hypertext. The C preprocessor is then used to deliver the final form of the hypertext page. 4. Discussion The primary benefit of user modeling for adapting hypertext is that it permits far greater customisation without additional complexity in the hyperspace the user sees. Our system operates on a large meta-hyperspace of the full range of customisability. Yet each user should see a quite manageable hyperspace. Two problems in this are related to user control and instability of the document. Both can be addressed. Simple approaches, like the viewable (and adjustable) user model permit straightforward control at that level. If the user wishes to explore the impact of their user model on hyperspace, they can do so by tinkering with the model and exploring the resultant hyperspace. Alternatively, we plan to provide tools to support such exploration by the user. This is particularly useful for establishing the language style that the user prefers. For example, it is easy for a user to read two versions of an explanation and say which they prefer. If we try to determine this in other ways, the user may need to be familiar with what we mean by terms such as ``abstract''. The second issue, of a changing hyperspace, is related. Our current plan is that as users progress through their learning of C, the material presented converges to the form of the printed text, and is intended for reference use. Tutorial level material would still be accessible deeper in the hyperspace. As the above description indicates, there is a merging of intelligent teaching systems and hypertext teaching materials. The user model that generates a customised hypertext is critical in building this bridge for customised communication. 5. References R Cook, J Kay (1993), `Tools for viewing um user models', SSRG Report 93/3/50.1, Dept of Computer Science, University of Sydney, Australia J Kay (1990), `um: a user modelling toolkit', Proceedings Second Intl User Modelling Workshop, Hawaii 1990. J Kay, `The um toolkit for reusable, long term user models', SSRG Report 94/3/36.2,