Structure-preserving knowledge-based system development through reusable libraries: a case study in diagnosis

Structure-preserving design has emerged as an important principle in Knowledge Engineering for developing knowledge-based systems. The principle prescribes the construction of increasingly more detailed models of the functionality of the knowledge-based system, while the information content and structure of the initial model are maintained throughout the process. It is, however, not trivial to come up with such models: their construction is still more an art than a science. Libraries with reusable components provide generic structures ready for selection and adaptation to the specific requirements at hand. We show the construction of a conceptual and formal model of a diagnostic reasoner through the use of two respective libraries, following the structure-preserving design principle. We include a proof that the high-level conceptual specification of the reasoner is ensured by the formal model presented. (C) 1997 Academic Press Limited.