Several knowledege based systems have been produced in the last decade, both in the industry and in research institutes. Some of these applications have been used in practical situations, but most of them have not been developed beyond the prototype stage. Such systems are usually desingned through an informal and empiric process which involves the implementation of a simple prototype of the system in the early phases of knowledge acquisition. This approach has beem widely criticized because it can make both the re-implementation and updating of knowledege bases cubersome, and seriously distort the way knowledge is represented. For this reason, a number of more formal methodologies for developing knowledge based systems have been proposed in recent years. Although such methodologies have provided some useful advice, none of them has yet proved to be applicable as whole in a wide range of situations. The first part of this paper consists of a general review of some of the existing formal methodologies. In the second part, the approach used for developing a knowledge based system for planning house building projects, named House Planner, is discussed.
Expert Systems; Knowledge Engineering; Artificial Inteligency