Towards refactoring of rule-based, in-place model transformation systems
Gabriele Taentzer, Thorsten Arendt, Claudia Ermel, Reiko Heckel · 2012
The more model transformations are applied in various application domains, the more questions about their quality arise. In this paper, we present a first approach towards improving the quality of endogenous in-place model transformation systems. This kind of model transformations is typically rule-based and well suited to perform model simulations and optimizations. After discussing suitable quality aims for this kind of model transformation systems and how they can be detected by smells, a first selection of refactorings is presented showing a variety of potential improvements of model transformation systems. Each refactoring is presented in a systematic way including an explanation how the quality is improved, a description of its pre- and post-conditions, a possible refactoring strategy, and an example. All discussed refactorings are implemented in Henshin, a model transformation engine based on graph transformation concepts, using Henshin in combination with the Eclipse plug-in EMF Refactor, a refactoring plug-in for defining and applying refactorings of EMF models.