Modeling Articulation Work in Software Engineering Processes

Abstract

Current software process modeling techniques do not generally support articulation work. Articulation work is the diagnosis, recovery and resumption of development activities that unexpectedly fail. It is an integral part of software process enactment since software processes can sometimes fail or breakdown. This paper presents a knowledge-based model of articulation work in software engineering processes. I t uses empirically-grounded heuristics t o address three problems in articulation work: diagnosing failed development activities, determining appropriate recovery, and resuming software processes. We first investigate the role and importance of articulation work with respect t o planned software development activities. We then outline a knowledge-based model of articulation work. The model has been implemented in a knowledgebased software process modeling environment called the Articulator. Combining the available software process modeling techniques and the model of articulation leads to a better foundation in process improvement and evolution.