1. Introduction
Industry and government organizations such as WallMart and General Motors increasingly adopt open source software (OSS) as alternative solution to proprietary commercial software products. Some well known OSS products, e.g., HTTP Server or Apache Tomcat, can be considered successful as they currently dominate their particular application domains as reported by Netcraft1 and The Serverside2 .
OSS projects can be viewed as extreme form of distributed software development based on criteria suggested by [4], such as: (a) highly distributed and volunteer contributors, (b) considerable cultural and time zone differences of contributor teams, (c) weak formal design, and (d) weak formal project and quality management. As software quality is believed to be strongly dependent on the quality of the software development process, quality assurance (QA) professionals and customers are particularly concerned about the quality of released OSS products and are interested in risk assessment of the associated OSS development processes. This background raises the need for investigating QA activities commonly performed by more or less successful OSS project communities in their development processes. From QA point of view, we can view Linus Torvalds’s [8] style of OSS development to produce high quality software, which encompasses the following practices: release early and often, delegate everything you can, be open to the point of promiscuity as a spawning ground for collaboration among developers and users to perform bug detection, bug assessment, bug tracking, and code peer review. In a commercial software project, in more semi-formal way some of these activities seem similar to inspection [2], which is an effective but also expensive approach to QA; however the community in OSS project has tackled the financial barrier as they build on self motivated volunteers.