Let's talk about bugs

Software testing is often a complex process potentially involving a large number of geographically distributed people with different perspectives and competencies. Software testers, software developers and project managers engage in discussions about the software errors found, they negotiate the relative importance of the bugs, they allocate responsibilities and resources, they coordinate who is doing what, etc. They talk about bugs. In order to coordinate and manage talking about bugs, a number of means for coordination are applied. The aim of this paper is to analyze coordination work in software testing in order to promote general requirements for computer support. We have studied the testing of more than 200,000 lines of code at Foss Electric, a Danish manufacturing company, and focused on two aspects: Firstly, the coordination activities related to the process of distributed registration, classification, diagnosis, correction, and verification of software errors, as well as the monitoring of the state-of-affairs of testing activities. Secondly, the mechanisms used to support the coordination. The analysis resulted in the identification of the need for computer support for coordinating this part of the software testing process, e.g., support of distributed classification, routing of information, and facilities providing an overview of state of affairs.