DESIGN AND EVALUATION OF AN ADAPTIVE ICON TOOLBAR

As information systems become increasingly important in many different domains, the potential to adapt them to individual users and their ne eds also becomes more important. Adaptive user interfaces offer many p ossible ways to adjust displays and improve procedures for a user's in dividual patterns of work. This paper describes an attempt to design a n adaptive user interface in a computer environment familiar to many u sers. According to one classification of adaptive user interfaces, the adaptive bar described in this paper would be classified as a user-co ntrolled self-adaptation system. At the user's convenience, the adapti ve bar offers suggestions for adding or removing command icons, based on the frequency and probability of specific commands. It also impleme nts these changes once the user has agreed to them. Beyond the adaptiv e bar, the general behavior of the whole user interface does not chang e, thereby allowing the user to maintain a clear general model of the system. This paper describes the decision-making algorithm implemented in the bar. It also describes the bar's self-adaptive behavior of dis playing the frequency of each icon's use through the icon's size. Fina lly, we present some encouraging preliminary results of evaluations by users.