Sandra de A. Siebra and Geber L. Ramalho
Universidade Federal de Pernambuco
Departamento de Informática
Caixa Postal 7851 - CEP:50732-970 - Recife - PE - Brasil
E-mail: (sas, glr)@di.ufpe.br
Computers are increasingly more pervasive in all sectors of daily life and, consequently, they are being used by people with a wide variety of backgrounds, preferences, abilities and goals. This state of affairs demands a significant research effort in interface design [ Preece 1994, Shneiderman 1992].
The emergence of adaptive interfaces [ Benyon 1993, Benyon & Murray 1988] provides a satisfactory framework for taking into account users heterogeneity. In adapting to the user's characteristics, these interfaces could, in principle, minimize training and support costs, as well as improve user's satisfaction and productiveness. However, so far, there is no general agreement on adaptive interface models. There is neither a widely accepted enumeration of relevant types of adaptations the interface should be able to undertake (not only cosmetic ones), nor a definite study on the impact of these adaptations in users' learning and performance.
Motivated on a five-year experience in dealing with users having different backgrounds, we have undergone a systematic study of the main humancomputer interaction problems. Once identified the potential helpful changes an adaptive interface should perform, we defined the guidelines for its development. Following, we proposed an adaptive interface model, called Athena, based on a distributed problem solving architecture. A prototype of Athena has been implemented in order to evaluate properly our initial working hypothesis.