Distribution of effort among software development artefacts: An initial case study

Abstract

Model-driven development aims at increasing productivity by raising the abstraction level of software specifications and introducing automated transformations for replacing lower level specifications. To assess benefits of replacing a legacy development process with a model-driven approach, one needs to establish a baseline of the current process with respect to the effort invested in the development artefacts. In this paper we report on an initial case study in which we investigate the main artefacts in the analysis and design phase with respect to required effort and perceived importance. We studied a non-model driven development of software based automotive functionality and our initial results show that a few artefacts receive the majority of effort, and that the artefacts that receive the most effort are not the most important ones. The initial results indicate that the distribution of effort between models and other artefacts is similar to that of model-driven projects in spite of the project being perceived and characterized as code-centric. © 2010 Springer-Verlag Berlin Heidelberg.