Dr.-Ing. Anthony Anjorin
Bidirectional Model-to-Text Transformations
Model-to-Text transformations play a central role in the context of Model-Driven Software Development (MDSD). Codegeneration and parser technologies have been researched for many years and not only theoretical results but also mature tool support is available. An open challenge is however catering for bidirectionality in this context, allowing possibly concurrent changes to models and text.
In this research topic, ideas from metamodelling and (triple) graph grammars are to be applied to tackle this challenge and theoretical results pertaining especially to bidirectionality are to be investigated and possibly transferred to the model-to-text case.
A feasible approach to detangle codegeneration/parsing from the actual transformation via a separation of a model-to-text transformation in a model-to-tree and a tree-to-text transformation has been investigated and presented in a workshop paper. Advantages include being able to use graph grammars (story-driven modelling, SDMs) and triple graph grammars (TGGs) for the model-to-tree transformation, thus providing support for bidirectionality.
Results have been integrated in the meta-CASE tool eMoflon (www.moflon.org) which was first of all reengineered and restructured to leverage state-of-the art model-driven technologies and tools. The architecture of the new system and reasons for reengineering have been presented in a paper for a workshop on the implementation of model-driven tools.