A usability-evaluation metric based on a soft-computing approach

Usability of software should measure both user preference and user performance. The notion of usability involves several dimensions. These dimensions include: system feedback, consistency, error prevention, performance/efficiency, user like/dislike, and error recovery. Each of these dimensions are characterized by fuzzy aspects and linguistic terms. This paper develops a model for each of the dimensions using fuzzy-set theory. It then uses the Takagi-Sugeno fuzzy-inference approach for developing an overall measure of usability. Results are presented for several different user interfaces.