A Formal Framework for Retainment Patterns for Trace-Based Model Transformations

Model-to-model (M2M) transformations play an important role within model-driven development. Modern M2M approaches support incremental updates to the target model according to changes in the source model(s). Bidirectional transformation approaches even allow to incrementally translate target model changes back to the source model. However, in some cases, e.g., if the target model should be refined either manually or automatically, it is important that target model changes are not overwritten if the original transformation is re-executed. There is currently only weak support for this kind of retainment by transformation engines. However, in many transformation engines a transformation trace is available which keeps record of a transformation's actions. In this paper, we exploit this information and define patterns which allow transformation engineers to trim transformations to facilitate the handling of target model changes. We describe a formal framework which serves as basis for implementing these retainment patterns.