Reliability Against the Odds : The Lessons of Open Source Communities for Developing Reliable Software

Open source software is increasingly used in mission-critical processes of organizations, signaling its acceptance as highly reliable software. That said, the development process in open source communities differs radically from conventional approaches to reliable software development. Stronger still, open source software seems to thrive under conditions of that are orthogonal to the conventional approach: extreme fragmentation, weak planning and coordination structures and lack of control over objectives, priorities and resource allocation. Therefore, we argue that open source communities provide us with valuable and innovative lessons for managing the production of reliable software - especially since the conditions for contemporary software development increasingly resemble those of open source communities rather than those of tightly controlled corporate environments. To draw out these lessons, this paper applies concepts from High Reliability Theory (HRT) on two recent case studies of open source communities, Linux and Apache. We identify a number of important reliability-enhancing mechanisms. Understanding these mechanisms is important to a) managers of corporate software developers participating in open source communities and b) managers who want to transfer mechanisms and lessons from open source communities to their organization