Inplace UML to UML Transformations: Constraint Generation and Partial Fixing for UML Models through Transformation
Working with complex models not only requires knowledge and skills to design them, but more importantly, change triggered to model elements by humans may violate the wellformedness or semantic design rules. Especially, semantic relationships within Unified Modeling Language (UML) models are complicated to preserve. Model-driven engineering standards such as the Object Constraint Language (OCL) support the necessary consistency checking. However, the manual generation of such constraint expressions and in advance the elimination of inconsistencies does not suit our needs. In this thesis, an incremental model manipulation approach (achieved through the ATLAS Transformation Language) is used to present an application for both traditional model transformation and constraint-driven modeling with the goal to eliminate existing inconsistencies (as far as possible). During the process of the transformation, constraints are generated to validate the model with - all in the context of the UML. In particular, 9 different constraint-driven scenarios, supporting class, sequence and statemachine diagrams/models, were developed. To illustrate a successful transformation process, the updated model is validated via the OCL subsequently. Model designers in the needs of consistency checking and constraint generation bene t from the automatic and incremental execution provided by the implementation.