Specification of Behaviour in Component Frameworks

Introduction There are currently few approaches for the specification of components and component frameworks that exist. Catalysis [DW98] which has a strong relationship to the UML (Unified Modelling Language) has gained public interest and seems to have reached the maturity to be applied in industrial component-based projects (e.g. COOL:Spex from Sterling Software). From an abstract viewpoint Catalysis defines an extensive method for the specification of components and component frameworks. Similar to the UML Catalysis emphasises the specification of functional and behavioural aspects of component software, but omits non-functional properties like timing. Catalysis is based on five key concepts for the specification of component software. Firstly, the definition of type models is supported. Here, Catalysis focuses on behavioural aspects specifying actions and invariants by the aid of OCL (Object Constraint Language) expressions. Secondly, collaborations of components describing how components in different roles interact can be specified. The concept of refinement for individual types as well as collaborations is used and the application of frameworks as a generic design or modelling unit. Finally, framework collaborations which represent a particular kind of framework utilising placeholder types and generalised actions that are used to support a flexible composition are also possible. On the other hand formal methods could be useful for the specification of component behaviour [TV98] and for proving that component frameworks meet their requirements. Concerning this aspect one result of the WCOP´98 was that there is still research required for quality assurance for component frameworks. We see in the formal specification and verification of components and component frameworks one approach to overcome this deficiency. cTLA(compositional TLA) is based on L. Lamports Temporal Logic of Actions [L94] and refers to the concepts of state transition systems, refinement mappings, and the separate definition of safety as well as liveness properties [HK94], [HK95]. Unlike TLA, the cTLA composition principle is oriented at CCS and Lotos and applies the principle of superimposition. In comparison with [AL93], [AL95] the cTLA processes do not interact via shared variables but perform joint actions. This stateless way of interaction has different benefits. Constraint-oriented processes can be especially represented [VSS88] that are well suited for the diagrams of the UML [GHK99]. Furthermore, the language supports decompositional proofs. A system is the logical conjunction of its processes and the language conventions assure the absence of contradictions in the system formula. Thus process properties are directly inherited to the system. The compositionality of …