Simplification of Semantically-Rich Model Transformations through Generated Transformation Blocks

This paper demonstrates a novel concept for the simplification of model transformations in which composite or complex objects are inserted into an existing model through a well-defined interface. The technique utilizes a model transformation from the domain of the modeling language into the domain of model transformation languages. The user specifies these semantically rich blocks using the original domain-specific modeling language. Then, a transformation generates the necessary model transformation graph to create an instance of the semantically rich, user-defined pattern. Users insert these generated patterns into their customized transformations. The approach is helpful for endogenous transformations in which existing objects may be refactored. It will also serve as a teaching tool for users who are unfamiliar with model transformations: specifically how to represent a newly-created model in the transformation domain. Finally, the approach is designed to reduce specification errors of model transformations in which new (semantically rich) blocks are inserted at key points, as the correctness of the semantically rich blocks is guaranteed, based on their construction in the original domain.