First international workshop on global integrated model management

Abstract

The aim of the First International Workshop on Global Integrated Model Management is to bring together researchers working on outstanding model driven engineering issues, with a view to discussing them, comparing their approaches, techniques or ideas, and letting these researchers take advantage of different perspectives on these issues. More particularly we are now see the second-generation modeling environments coming up, where the focus is no more on one general purpose modeling language (like UML 2.1 for example), but on a large family of more or less related domain specific languages (DSLs) corresponding to different corporations, different needs, different levels of abstraction, etc. First generation tools were mainly based on one metamodel (e.g. UML) while most recent tools are metamodel agnostic and able to deal with various kinds of metamodels or DSLs. The advantages of using a set of DSLs instead of one General Purpose Language (GPL) are easy to understand, but we also need to cope with the corresponding drawbacks. The resulting fragmentation may be addressed by the existence of a metametamodel (like OMG's MOF) to which all DSLs may conform. They are however various proposal for this unique metametamodel (e.g. EMF ECORE, Microsoft DSL/Tools, etc.). Another solution is to use precisely defined and semantically sound model transformation and composition operators between the various DSLs. However these solutions alone are not sufficient to handle the complex situation arising from the use of a high number of artifacts of different types like models, metamodels, model transformations, model weaving, etc. This is true when various technical spaces [1] as well as different domains are taken into account. The notion of metamodel takes different incarnations in each technical space (e.g. grammars in Grammarware, ontologies in Ontologyware, schemas in XMLware, etc.). The fragmentation problem remain however the same. Therefore we need to devise general and solid solutions to global model management in situations where the various artifacts are distributed as general resources available in a web based distributed system. As a consequence global model management is bringing a set of new research challenges that we have to address on the short to medium term. To solve these open problems, it will be interesting to apply the same kind of methods that have originated them, i.e. model engineering solutions and consequently we need to invent these new integrated methods to global model management. These new engineering techniques need to be based on a solid modeling theory, dealing wth composition, refinement, abstraction and other useful semantic based development relationships between models.