Ensuring Containment Constraints in Graph-based Model Transformation Approaches

Abstract

Within model driven software development, model transformation has become a key activity. A number of transformation approaches for metamodel-defined modeling languages have been developed in the past years and are going to be established in research and industry. None of these have made it to a standard yet. There is a demand for correct model transformation in various senses. Formal methods are helpful for showing correctness issues of model transformations. As one approach, graph transformation has been applied to the field of model transformation and is a perspective for achieving provable correct model transformations. We show in this paper, that containment associations as proposed by the OMG are an integral part of MOF-based languages and imply a couple of constraints which must be ensured in model transformation approaches. Based on a double-pushout approach to graph transformation, conditions are stated that ensure these containment constraints. This is an important step for achieving formal transformation semantics for modeling languages based on MOF, or specifically EMF.