On Integrating Triple Graph Grammars and OCL for Model-Driven Development

Software systems become more and more complex. Despite significant advances in code-centric technologies such as advanced programming languages and integrated development environments (IDEs), developing complex software systems is still a laborious task. Model-centric software development has emerged as a promising paradigm, indicating a shift from "code-centric" to "model-centric" engineering. This paradigm puts forward a necessity as well as a challenge of a formal foundation for presenting precisely models and supporting automatic model manipulations. This thesis focuses on a model-driven approach in which metamodeling and model transformation are seen as the core of the approach. On the one hand, metamodeling, which is a means for defining modeling languages, aims to precisely present models for specific purposes. On the other hand, model transformation is a robust approach for (1)~transforming models from one language into another language, (2)~tackling the challenge how to choose a right level of abstraction, and to relate levels of abstraction with each other, and (3)~supporting software evolution and maintenance. We propose an integration of two light-weight formal methods, the Object Constraint Language (OCL) and Triple Graph Grammars (TGGs), in order to form a core foundation for the model-driven approach. TGGs incorporating OCL allow us to explain relationships between models in precise way. On this foundation, we develop a textual language in order to describe transformations and to realize transformations. From such a declarative description of transformations, we propose employing OCL in order to derive operational scenarios for model transformation. This results in an OCL-based framework for model transformation. This framework offers means for transformation quality assurance such as verifying model transformations in an on-the-fly way, checking well-formedness of models, and maintaining model consistency. We explore case studies showing the practical applicability of our approach. Especially, TGGs incorporating OCL allow us to describe the operation semantics of use cases in a precise way. This approach on the one hand can be broaden in order to describe the operational semantics of modeling languages. On the other hand, it allows us to generate scenarios as test cases, to validate system behavior, and to check the conformance between use case models and design models. This supports basic constructions of an automatic and semi-automatic design.