Bridging formal models : an engineering perspective

Abstract

The thesis presents different techniques that can be used to build formal behavioral models. If modal properties are formulated, the models can be subjected to verification techniques to determine whether a model possesses the desired properties. However many native environments do not facilitate tools or techniques to verify them. Hence, these models need to be transformed into other models that provide suitable techniques for a formal analysis. The transformations are classified into two engineering approaches, namely syntactically engineered models and semantically engineered models. Syntactically engineered models are constructed from input specifications without explicitly considering the semantics. Semantically engineered models are constructed from input specifications by explicitly considering the semantics. The syntactic engineering approach presents four dedicated modeling techniques that construct or disseminate verification results for formal models. The first modeling technique describes a way to create models from system descriptions that specify concurrent behavior. Here, we model three variations of a 2×2 switch, for which the models are subsequently compared to models created in the specification languages: TLA+, Bluespec, Statecharts, and ACP. The comparison validates that mCRL2 is a suitable specification language to model descriptions or specify the behavior for prototype systems. The second syntactic technique constructs an mCRL2 model from a software implementation that operates a printer for printing Printed Circuit Boards. The model is used to advise (other) software engineers on dangerous language constructs in the control software. Hence, the model is model checked for various safety properties. The implementation is modeled through an over-approximation on the behavior by abstracting from program variables, such that only interface calls between processes and non-deterministic choices in procedures remain. The third modeling technique describes a language transformation from the language Chi 2.0 language to the mCRL2 language. The purpose of the transformation is to facilitate model checking techniques to the discrete part of the Chi 2.0 language.