Formally refining software architectures with π-ARL

Abstract

π-ARL is a formal (executable) architecture refinement language providing architecture-centric refinement primitives and constructs for their compositions. When applied, refinement actions expressed in π-ARL refine architectural models described in π-ADL outputting new refined models described in π-ADL. Enabling stepwise architecture refinement is a new challenge for the formal development of complex software systems. This article presents a simple but realistic case study of the use of π-ARL for architecture refinement. It illustrates the expressiveness and usefulness of π-ARL. The case study addresses the modelling and refinement of the software architecture of a Data Acquisition System. It covers a simple, yet frequent, architecture refinement that would be problematic for most other refinement techniques. Several refinement steps are performed, each dealing with a simple refinement, in order to achieve a concrete architecture.