The epistemology of validation and verification testing

Abstract

We wish to be able to give formal definitions (in the sense of science or engineering) for concepts like requirements validation and for the relationship between a requirements specification and an abstract design of the intended system. Ditto validation of designs and the final executable application with respect to the original "application concept", on the one hand, and the requirement specification, on the other. We have been developing a framework based on the work of the logical empiricists and other analytic philosophers over the last 80 years to support our understanding of software engineering concepts. Recent developments (dating from the 80s)in the area of "confirmation" (of a hypothesis concerning a theory by some (experimental) evidence) promises to illuminate some of these problematic concepts. In this talk we address the problem of establishing the very relation between requirement specifications and scenarios, as used, for example, in UML. The same framework can also be applied to the problem of testing implementations against designs, so called verification testing. © IFIP 2005.