Abstract

There is a large semantic gap between the level at which a solution to a problem is conceived by a domain expert or a requirements engineer and the level at which it is programmed on a computer. Traditional approaches to minimizing the impact of this gap have been based on using modeling languages and/or domain specific languages that have clear formal semantics. Designing the modeling or domain specific language is not enough, as many problems still remain. These problems pertain to being able to rapidly develop an implementation infrastructure for the language in a provably correct manner and being able to interoperate between multiple languages proposed for a given domain. In this dissertation we propose and develop a semantics-based framework---founded on Horn Logical Denotational Semantics---to solve these problems. We show how our approach can be used for rapidly obtaining the implementation infrastructure (interpreter, compiler, debugger, and profiler) for domain specific languages in a provably correct manner, as well as for working with multiple domain specific language in an interoperable manner. We also show how our approach works even in presence of unstructured control constructs.