Validation and Verification of KADS Data and Domain Knowledge

Abstract The VITAL project aims at building a methodology and its environment to support the cost, time-effective, and quality-driven development of knowledge-based systems (KBS). In this context verification and validation issues have a central role. This paper shows how VITAL adds a set of formalisms to KADS and adopts the Slage's approach to knowledge specification to enable verification and validation of expertise models. In particular if formalization is now the common way to deal with knowledge specifications verification, other techniques and other formalisms are required to cope with their validation. Tables are a very suitable representation to bring knowledge engineers and experts together to compare their interpretation of the expertise and ensure that the implemented system is close to the way the expert solues the problem. And constraint programming is an appropriate technique for data and domain knowledge validation.