The Social Order of Open Source Software Production

Introduction

This article contributes to the sociological understanding of open source software (OSS) production by identifying the social mechanism that creates social order in OSS communities. The concept of social order is used here in its most basic sense as describing a situation in which actors have adjusted their actions to each other. This order is indeed very high in OSS communities, who produce large and highly complex software from many independent contributions. It is even astonishingly high when we take into account how few of the most common tools for creating order—rules, commands, and negotiations—are used. Therefore, most analysts agree that OSS is produced in a distinct “new” mode that is qualitatively different from the “corporate way” of software production.

However, none of the four strands of literature on OSS production has produced a consistent explanation of the way in which this amazing order is achieved. The participant-observer literature has proposed metaphors that emphasize the decentralized, democratic, open, and communal nature of OSS, notably the “cooking pot market” (Ghosh, 1998) and the “bazaar” (Raymond, 1999). These metaphors, while suggestive, are not grounded in social theory. Economics is still fascinated by the voluntary contributions to a public good, and has consequently focused on motivations to contribute to OSS (Dalle & Jullien, 2003; Lerner & Tirole, 2002; von Hippel & von Krogh, 2003). However, neither these investigations nor the generalized questions about transaction costs (Demil & Lecocq, 2003) or about the allocation of efforts to modules (Dalle, David, Ghosh, & Steinmueller, 2004) capture the specific ways in which an ill-defined group of people manages to produce a complex good. These ways have been looked at primarily in the context of management and software engineering analyses, which produced interesting case studies of the coordination of individual OSS projects such as Linux, Apache, Perl, Sendmail, Mozilla, and others (Holck & Jørgensen, 2005; Iannacci, 2003; Jørgensen, 2001; Koch & Schneider, 2002; Lanzara & Morner, 2003; Mockus, Fielding, & Herbsleb, 2002). Some analysts tried to compare OSS communities to “traditional organizations” (Sharma, Sugumeran, & Rajagopalan, 2002) or to catch the specific mode of OSS production with generalized concepts such as “virtual organization” (Gallivan, 2001) or “distributed collective practice” (Gasser & Ripoche, 2003). However, these concepts are similar to the metaphors in the observer-participant literature in that they are ad hoc generalizations that are not embedded in theories of social order or of collective production. Finally, sociological analyses have contributed the idea of a gift economy (Bergquist & Ljungberg, 2001; Zeitlyn, 2003), various concepts of community (Edwards, 2001), social movements (Hess, 2005; Holtgrewe & Werle, 2001), the hacker culture (Lin, in this volume), and applications of actor-network theory (Tuomi, 2001). These sociological accounts focus on the specificity of social relations in OSS communities and more or less entirely disregard the specific mode of production employed by these communities.