The influence of domain-specific abstraction on evolvability of software architectures for information systems

Abstract

Our research focuses on how evolvable conceptual models of software architectures can be built. We study the earliest steps in the modelling process, when initial decisions are taken about which concepts should be modelled in which way. We feel that an important factor in the evolvability of a conceptual model is the type of domain-specific abstraction that is used to model these concepts. This point-of-view raises questions about which type and level of abstraction should be used at which places in a software architecture in order to achieve a certain type of evolvability.