Feature Interaction in Composed Systems

Abstract

The history of computer science has shown that decomposing software applications helps managing their complexity and facilitates reuse, but also bears challenging problems still unsolved, such as the assembly of the decomposed features when non-trivial feature interactions are involved. Examples of features include concerns or aspects, black box or white box components, and functional and non-functional requirements. Approaches such as object-oriented and component-based software development, as well as relatively new directions such as aspect-oriented programming, multi-dimensional separation of concerns and generative programming, all provide technical support for the definition and syntactical assembly of features, but fall short on the semantic level, for example in spotting meaningless or even faulty combinations. At previous ECOOPs, OOPSLAs and GCSEs dedicated events have been organised around the aforementioned technologies, where we experienced a growing awareness of this feature interaction problem. However, feature interaction is often merely dismissed as a secondary problem, percolating as an afterthought while other issues are being addressed. The intention of this workshop was to be the first co-ordinated effort to address the general problem of feature interaction in composed systems separately from other issues.