Regresar

An extensible meta-modelling assistant

Abstract:

Meta-models play a pivotal role in Model-Driven Engineering (MDE). They are used to create domain-specific models, and to type model management operations like model transformations or code generators. However, even though creating meta-models is a common activity, it is currently mostly a manual activity, which does not profit from existing knowledge. In order to facilitate the meta-modelling task, we propose an extensible meta-modelling assistant. While primarily focussed on helping in the creation of meta- models, it can also help in creating models. The assistant permits the provision of heterogeneous data description sources (like ontologies, RDF data, XML schemas, database schemas and meta-models), and enables their uniform querying. Different kinds of queries are supported, and improved through synonym search. Query results are prioritized through sense disambiguation, can be graphically visualized, and incorporated into the (meta-)model being built. The assistant has been realized within Eclipse, and its architecture has been designed to be independent of the meta-modelling technology used. As a proof- of-concept, we show its integration within DSL-tao, a pattern-based meta-modelling tool built by our group, and two other tools developed by third-parties. The usefulness of the system is illustrated with a running example in the process modelling domain.