A Generic Technique for Domain-Specific Visual Language Model Refactoring to Patterns

Abstract

As the popularity of domain-specific visual languages (DSVLs) grows, concerns have arisen regarding quality assurance and evolvability of their meta-models and model instances. In this paper we address aspects of automated DSVL model instance modification for quality improvement based on refactoring specifications. We propose a graph transformation-based visual language approach for DSVL authors to specify the matching and discovery of DSVL “bad model smells” and the application of pattern-based solutions in a DSVL meta-tool. As an outcome, DSVL users are provided with pattern-based design evolution support as refactorings for their DSVL-based domain models.