Bridging, Patching and Keeping the Work Flowing: Defect Resolution in Distributed Software Development

Abstract

This paper reports on results from a long-term field study of globally distributed software development projects within a multinational organization. The research explores the issues involved in performing global software development, utilizing a perspective informed by CSCW research concerning the local organization of work practices and the key role of workers in being able to intervene in the ‘flow of work’ where necessary. The paper also raises some more general questions concerning the field of Global Software Development (GSD), in terms of the concepts and methods being used in the area. Our contribution is in the form of a CSCW-informed empirical study of the use of defect (or ‘bug’) tracking systems—systems which support the identification, classification and resolution of defects in the emerging software. In one case, the team persisted with a defect tracking system that they had used for years and maintained it in parallel with a system used by co-workers in other countries—all the while attempting to implement a bridge between the two. In the other, we report on how local software patches were created to allow for local work to proceed while not interfering with the existing coordination mechanisms between the local site and remote co-workers who were responsible for creating daily builds according to the overall project plan. In both cases, local practices were shaped by the necessity to keep work flowing across the whole project, even if this involved what might, at first sight, seem to go against project-wide practice. We discuss implications of these findings in terms of a key distinction between externally-prescribed ‘workflow’ and internally-managed ‘flow of work’ activities. We also explore how a heterogeneous ‘assembly’ of variably coupled systems may be the most appropriate image for technological support of distributed teams as they keep the work flowing in an orderly fashion. Overall, our work suggests that studies of global software development can profit from the CSCW tradition of workplace studies both conceptually and methodologically.