LORIA, UMR 7503 (CNRS, INRIA, Universités de Nancy) Campus Scientifique, BP 239, F54506 Vandœuvre-lès-Nancy Cedex, France tel: +33 (0)3 83 59 20 32, fax: +33 (0)3 83 41 30 79 e-mail: Noelle.Carbonell@loria.fr
The advent of a world wide "information society", which can be viewed as the next major social evolution, may increase social exclusion, unless user interface designers consider accessibility issues carefully, and take into account emerging guidelines and standards ( Stephanidis et al. 1997, HFES/ANSI 1997, World Wide Web Consortium 1998). Achieving universal accessibility raises crucial issues indeed, which have not yet been solved, namely meeting the usability ( Nielsen 1993) requirements (i) of all standard user categories (i.e. novices, occasional and expert users), (ii) in all contexts of use, especially emerging ones (home automation, online services for the general public, mobile computing, etc); and (iii) providing all users including disabled people with appropriate human-computer interaction (HCI) facilities.
The concepts of adaptability and adaptivity have been proposed to solve issues (i) and (iii). In particular, they have been implemented successfully in the European Project AVANTI ( Stephanidis et al. 1998). However, both concepts present limitations: Adaptations are predefined (static), and their scope is limited. In addition, current implementations of adaptivity are still too crude and unreliable to be useful, in-as-much as dynamic self-adaptation is still based mostly on the analysis of the user's actions, which provide insufficient information on capabilities, preferences, goals and strategies; moreover, notifying users of this evolution, and enabling them to control it, is difficult and necessitates long-term research efforts.