Validating Transformations for Semantic Anchoring.

Making Domain-Specific Modeling Languages a part of a tool chain, a part of a proven development process, or the subject of verification cannot be achieved without the precise specification of the language and the models expressed in it. Defining formal semantics for modeling languages is a widely accepted solution to this problem. We have developed methods, techniques and processes to provide a systematic mapping - which we call semantic anchoring - that support the scaling of these formal definitions to large modeling languages. Although a semantic mapping is a definition and behaves as a set of axioms for formal verification, we argue that a semantic mapping can conflict with the informal intentions of the language designer, resulting in a counterintuitive DSML, and should therefore be validated. This paper proposes a solution that involves formalizing the language designer's intentions about the semantic mapping and validating the consistency between the two by applying model finding techniques.