Integrating semantic reasoning in information loss-based transformation chain rankers

Models transformations are at the heart of model-driven engineering. They are increasingly recognized as crucial entities to achieve superior automation in many software engineering areas, whether it be requirements traceability, consistency restoration, and model management. As with many knowledge-intensive artifacts, model transformations can be challenging to design, develop, and maintain. Thus, defining complex model transformations by chaining existing ones is key to enhanced quality and increased reuse. Identifying the right transformation chains demand dedicated support when multiple paths are available to bridge a source metamodel with a target one. Metamodel coverage and Information Loss are among already established factors that can be adopted for supporting chain selections. In this work, we introduce the notion of Semantic Importance for metamodel elements involved in the transformation chains under analysis. The goal is improving the estimation accuracy of the Information Loss, which is being considered for ranking the possible transformations chains. The approach is supported by CITRIC+ tool, which includes a semantic reasoner able to select chains that induce the lowest Information Loss, with respect to the Semantic Importance specified by modelers, using a dedicated DSL.