Existing requirements engineering approaches manage broadly scoped requirements and constraints in a fashion that is largely two-dimensional, where functional requirements serve as the base decomposition with non-functional requirements cutting across them. Therefore, crosscutting functional requirements are not effectively handled. This in turn leads to architecture trade-offs being mainly guided by the non-functional requirements, so that the system quality attributes can be satisfied. In this paper, we propose a uniform treatment of concerns at the requirements engineering level, regardless of their functional, non-functional or crosscutting nature. Our approach is based on the observation that concerns in a system are, in fact, a subset, and concrete realisations, of abstract concerns in a meta concern space. One can delineate requirements according to these abstract concerns to derive more system-specific, concrete concerns. We introduce the notion of a compositional intersection, which allows us to choose appropriate sets of concerns in our multi-dimensional separation as a basis to observe trade-offs among other concerns. This provides a rigorous analysis of requirements-level trade-offs as well as important insights into various architectural choices available to satisfy a particular functional or non-functional concern.
This paper brings together aspect-oriented and multi-perspective approaches to requirements engineering to tackle the tyranny of dominant decomposition during requirements analysis. It drives a shift away from the currently dominant philosophy in requirements engineering whereby functional and non-functional requirements are analysed in a two-dimensional fashion. Instead, it focuses on a multi-dimensional view where the relationships and dependencies amongst concerns are studied from multiple perspectives and these perspectives projected together through composition specifications to formulate a global understanding. It was accepted in a highly competitive conference programme (acceptance rate: 20%) and has already received 28 citations according to Google Scholar. RAE_import_type : Conference contribution RAE_uoa_type : Computer Science and Informatics