Collaborative Modeling Among Cross-Domain Teams

点击翻译按钮获取中文摘要

Complex systems engineering typically involves many stakeholder groups working in a coordinated manner on different aspects of a system. Each aspect addresses a specific set of system concerns and is associated with a domain space consisting of problem or solution concepts described using specialized terminology. A domain-specific modeling language (DSL) allows engineers in a stakeholder group to address their specific system development concerns by building and using models that express abstractions in domain-specific terms. Complex heterogeneous systems engineered by heterogeneous teams thus require that multiple DSLs are used cooperatively to support design, execution, and validation of the modeled system.******The field of model-driven engineering (MDE) provides a promising foundation to support modeling, executing, and validating domain-specific models. However, the lack of focus on the collaboration and integration of models and languages among teams and users has been a major impeachment for true industrial adoption, despite some recent success stories reported.******There are many fundamental challenges to solve in order to enable various stakeholders to work collaboratively on a common modeled system. In this program, we will concentrate our efforts to address scalability and usability issues in such distributed environments. The long-term goal of this research is to produce a framework where engineers and domain-experts can collaboratively model and synthesize complex heterogeneous systems.******We will investigate the appropriate storage representation and medium for a model repository able to handle a large amount of large-scale models (in the order of millions models containing millions of elements). Executing such models requires an imperative improvement in the performance of model transformations, by taking advantage of concurrency and the distributed resources available in the framework. Because multiple users will be managing models by means of multiple views on a same model, the framework must ensure a synchronization to keep models and views consistent with each other. Furthermore, each model or view will evolve over time during the development process and thus we will investigate for a domain-specific version control system that allows differencing and merging mechanisms adapted to the domain of the model, by taking into account the semantics and concrete syntax of models. Finally, we believe that the workspace of the domain-specific user must give him a sensation of complete immersion in his domain. Therefore, the framework must support the automatic generation of domain-specific IDEs that are customizable per domain and per user preference.