On the Need for Artifact Models in Model-Driven Systems Engineering Projects

Model-driven development has shown to facilitate systems engineering. It employs automated transformation of heterogeneous models into source code artifacts for software products, their testing, and deployment. To this effect, model-driven processes comprise several activities, including parsing, model checking, generating, compiling, testing, and packaging. During this, a multitude of artifacts of different kinds are involved that are related to each other in various ways. The complexity and number of these relations aggravates development, maintenance, and evolution of model-driven systems engineering (MDSE). For future MDSE challenges, such as the development of collaborative cyber-physical systems for automated driving or Industry 4.0, the understanding of these relations must scale with the participating domains, stakeholders, and modeling techniques. We motivate the need for understanding these relations between artifacts of MDSE processes, sketch a vision of formalizing these using artifact models, and present challenges towards it.