Knowlege elicitation in design

Abstract

The lack of intelligent support offered by current Computer Aided Design (CAD) systems, despite the obvious need, has resulted in attempts to apply proven methods of KBS (Knowledge Based Systems) and AI (Artificial Intelligence) to the field of CAD, in order to bridge the intelligence gap. Experience gained from the application of KBS technology suggests that the incorporation of higher levels of knowledge will result in more powerful, usable and intelligent computer systems. The power of such intelligent systems will clearly depend on the quality and depth of knowledge on which they are based. It is argued that the most direct and accurate source of such information is provided by the domain expert. However, the transference of expert understanding remains the difficult problem of knowledge elicitation. This paper argues that thorough knowledge elicitation procedures in the domain of design are needed to enable the development of more powerful ICAD systems. An important concept arising from the use of KBS technology is the value of a human communicable understanding of the domain, in terms of a testable, machine independent model of cognition. In this paper we describe a method of knowledge elicitation that has been applied in the domain of page layout design. Using the results of this method an elaborate model of the page layout design has been successfully developed. It is argued that this study demonstrates the feasibility and usefulness of applying knowledge acquisition techniques to design tasks.