Engineering Model Recommender Foundations - From Class Completion to Model Recommendations

Reuse has been widely carried out successfully, but not with models in Model Driven Engineering. Reasons seem manifold and conceptual issues and poor tool support are among them. A closer look at the tools available shows that models are often held in repositories which merely exceed versioning and indexing support. But model reuse requires mature approaches and tool support to become successful. We created a solid conceptual foundation and found recommendations as one solution, which in turn need appropriate data. We engineer these data subsequently and explain our design rationales. In a nutshell, we create a knowledge library comprising of elements which are connected on generic, semantic, and syntactic level. This knowledge library forms an enhanced knowledge graph enabling chain recommendations.