How Negotiation Shapes Coordination in Distributed Software Problem Management

All software exhibits operational problems as well as opportunities for redesign. How software projects handle problems and ongoing redesign efforts – Software Problem Management or "SWPM" – is not well understood, especially in open distributed communities such as Free/Open Source Software (F/OSS) projects. Little is known about the factors that influence performance or outcomes (e.g. time, cost, quality, sustainability) of software development communities involved in SWPM. Publicly accessible repositories of SWPM data, such as bug report repositories maintained by F/OSS development communities, provide vast stores of data about distributed SWPM practices and knowledge. Individual bug reports (BRs) within these repositories are texts that provide detailed records of community activities including responses to, and analyses of, problems. By examining these records we can develop detailed pictures of how many collaborative, distributed, computermediated SWPM processes unfold, creating finer-grained accounts of how coordination actually works, and extending current theories of computer-supported collaboration. This paper reports on several ways in which activity and information interact in SWPM. We have identified a number of "basic social processes" (BSPs) that shape how participants organize and use information, build and maintain community knowledge, and coordinate community activities at a very large scale (e.g., several gigabytes of online information, tens of thousands of reporters, over 50,000 simultaneously open problems). We focus on how one BSP, negotiation, shapes coordination activities in SWPM. Using detailed qualitative methods, we report on the varieties and frequencies of negotiation practices, and analyze how different types of negotiation in different situations influence what knowledge is available to participants, how collective activity gets focused in specific directions, and how the organization of information shapes activity. Overall, this provides a more detailed understanding of key collaborative distributed coordination mechanisms than has been available in the CSCW literature to date.