DSL/Model Co-Evolution in Industrial EMF-Based MDSE Ecosystems

Model Driven Engineering and Domain Specific Languages (DSLs) are being used in industry to increase productivity, and enable novel techniques like virtual prototyping. Using domain specific languages, engineers can model a systems in terms of their domain, rather than encoding it in general purpose concepts, like those offered by UML. However, domain specific languages evolve over time, often in a non-backwards-compatible way with respect to their models. When this happens, for models to remain usable they need to be co-evolved. Because the number of models in an industrial setting grows so large, manual co-evolution is becoming unfeasible calling for an automated approach. Many approaches for automated co-evolution of models exist in literature, each operating in a highly specialized context. In this paper, we present a high-level architecture that tries to capture the general process needed for automated co-evolution of models in response to DSL evolution. We then assess to what extent existing approaches already implement this architecture, and conclude which challenges are still open.Model Driven Engineering and Domain Specific Languages(DSLs) are being used in industry to increase productivity,and enable novel techniques like virtual prototyping. UsingDSLs, engineers can model a systems in terms of their domain,rather than encoding it in general purpose concepts,like those offered by UML. However, DSLs evolve over time,often in a non-backwards-compatible way with respect totheir models. When this happens, models need to be coevolvedto remain usable. Because the number of modelsin an industrial setting grows so large, manual co-evolutionis becoming unfeasible calling for an automated approach.Many approaches for automated co-evolution of models withrespect to their DSLs exist in literature, each operating in ahighly specialized context. In this paper, we present a highlevelarchitecture that tries to capture the general processneeded for automated co-evolution of models in response toDSL evolution, and assess which challenges are still ope