Fast Prototyping of Expert Systems Using Prolog

Publisher Summary This chapter focuses on fast prototyping of expert systems using Prolog. Prolog is a powerful language for expert systems programming. Its power comes mainly from its data structuring mechanism, built-in pattern matching, and other symbolic programming features. For prototyping to achieve its goal of exposing the potential user to a working system behavior, it is important that a prototype be developed rapidly and cheaply. The practical limitation of Prolog can turn out to be its relative computational inefficiency, relative compared with procedural languages, such as Pascal and C. In addition, not many existing Prolog implementations convincingly provide sophisticated user interaction facilities, graphics, and links for smooth and easy connection to other programming languages. As efficiency and these additional features are less important in prototyping in general, they are less critical in prototyping of expert systems as well. On the other hand, the expressive power of Prolog is such that developing a prototype in Prolog typically progresses simultaneously with the general development of ideas and decisions regarding knowledge representation and inference mechanisms. Although powerful, sophisticated knowledge engineering environments typically take long time to learn to use, are more expensive and can be computationally inefficient. A significant and often overlooked drawback of such environments is simply the fact that there is less programming culture built around each particular environment than around general programming languages such as Prolog.