A Solution to the Flowgraphs Case Study using Triple Graph Grammars and eMoflon

Abstract

After 20 years of Triple Graph Grammars (TGGs) and numerous actively maintained implementations, there is now a need for challenging examples and success stories to show that TGGs can be used for real-world bidirectional model transformations. Our primary goal in recent years has been to increase the expressiveness of TGGs by providing a set of pragmatic features that allow a controlled fallback to programmed graph transformations and Java. Based on the Flowgraphs case study [7] of the Transformation Tool Contest (TTC 2013), we present (i) attribute constraints used to express complex bidirectional attribute manipulation, (ii) binding expressions for specifying arbitrary context relationships, and (iii) post-processing methods as a black box extension for TGG rules. In each case, we discuss the enabled trade-off between guaranteed formal properties and expressiveness. Our solution, implemented with eMoflon (www.emoflon.org) our metamodelling and model transformation tool, is available as a virtual machine hosted on Share [15].