A Method for Aspect-oriented Meta-Model Evolution

Abstract

Long-living systems face many modifications and extensions over time due to changing technology and requirements. This causes changes in the models reflecting the systems, and subsequently in the underlying meta-models, as their structure and semantics are adapted to adhere these changes. Modifying meta-models requires adaptations in all tools realizing their semantics. This is a costly endeavor, especially for complex meta-models. To solve this problem we propose a method to construct and refactor meta-models to be concise and focused on a small set of concerns. The method results in simpler metamodel modification scenarios and fewer modifications, as new concerns and aspects are encapsulated in separate meta-models. Furthermore, we define design patterns based on the different roles meta-models play in software. Thus, we keep large and complex modeling projects manageable due to the improved adaptability of their meta-model basis.