Model-Driven Co-evolution for Agile Development

In agile Model-Driven Engineering, the evolution of diverse software artifacts is omnipresent. Especially the evolution of metamodels, defining the grammar of, e.g., Domain Specific Languages, is quite challenging, since many dependent artifacts, including models and transformations, have to be co-evolved to re-establish consistency. Although much research effort has been spent to automate co-evolution, crucial issues remain open and a systematic survey of the state of research across different domains is still missing. This paper provides an extensive survey evaluating various co-evolution approaches also from areas in software engineering like data, ontology, and grammar engineering on basis of a detailed set of criteria serving as a research roadmap for further developments in the area of co-evolution for agile MDE. Based on these results, a conceptual co-evolution framework is presented and illustrated by a running example especially targeting a decrease in co-evolution effort, an increase in co-evolution consistency, and an advance in extensibility.