Synthesis and Analysis of Component Connectors (SYANCO)

点击翻译按钮获取中文摘要

SYANCO is concerned with the development of a universal framework to support the synthesis and analysis of complex systems composed of heterogeneous components at various levels of abstraction. This involves tools and techniques for design and validation of (1) individual components, (2) individual pieces of their connecting glue code, and (3) the composition of components and their glue code. Contemporary well-known tools and techniques for automating and/or assisting design and validation of software are best suited in the development of individual components. They become inadequate for combining coarse-grain components, precisely because design and validation of the glue code to combine disparate components into a coherent system, are less understood. In SYANCO, we will work on the last two issues, above, emphasizing compositionality: we will adapt, extend, and/or develop new models, tools, and techniques for compositional reasoning about, as well as modeling, analysis, and construction of component-based systems, by primarily focusing on their connector glue code. Typically, programming models, languages, and systems use variations of function calls, method invocation, remote procedure calls, inheritance, delegation, etc., for software composition. Operators for exogenous composition, i.e., composition from outside, of two pieces of software by a third-party are ostensibly rare. Yet, composing a system out of heterogeneous third-party components requires precisely operators that not only interconnect and facilitate their intercommunications, but also coordinate their individual activities into a coherently orchestrated whole. In fact, most of the central issues in component composition involve coordination. Reo is a exogenous coordination language based on a calculus of channel composition [Arb04]. Reo has attracted interest and attention because it offers a powerful glue language for compositional construction of connectors that combine component instances into a system and orchestrate their mutual interactions. Since early 2003, the German and Dutch teams in SYANCO have had a continuing close collaboration on Reo and formal models for compositional verification of connector circuits and reasoning about them [ABRS04, ABdBR04, ABdB+05, BSAR05, ABC+05]. Building on this productive history, SYANCO (i) creates a formal context for our on-going bilateral research which combines complementary expertise on component-based software architectures and coordination models (CWI) and model checking and quantitative analysis of complex systems (Bonn), on the basis of common expertise on semantics for concurrency, (ii) provides funding to support mutual visits by the members of the two teams, as well as international visitors; and (iii) directly supports the expanding activities of the German partner into actual tool building by funding two new position in their team.