Combining architectural design decisions and legacy system evolution

Abstract

Software development is characterized by ongoing design decisions that must take into account numerous requirements, goals, and constraints. When changing long-living and legacy systems, former decisions have to be considered. In order to minimize the risk of taking wrong or misleading decisions an explicit representation of the relevant aspects is crucial. Architectural decision modeling has proven to be an effective means to represent these aspects, the required knowledge, and properties of a potential solution. However, existing approaches do not sufficiently cover the ongoing evolution of decisions and artifacts. They fail in particular to represent relations to existing systems on a fine-grained level to allow for impact analysis and a later comprehension of decisions. Furthermore, the effort for capturing and modeling of design decisions has to be reduced. In our paper we integrate existing approaches for software architectural design decision making. We extend them by fine-grained traceability to elements of existing systems and explicit means for modeling the evolution of decisions. We show how relevant decisions can easily be identified and developers are supported in decision making. © 2014 Springer International Publishing Switzerland.