We propose a new model of knowledge creation in purposeful, loosely-coordinated, distributed systems, as an alternative to a firm-based one. Specifically, using the case of Linux kernel development project, we build a model of community-based, evolutionary knowledge creation to study how thousands of talented volunteers, dispersed across organizational and geographical boundaries, collaborate via the Internet to produce a knowledge-intensive, innovative product of high quality. By comparing and contrasting the Linux model with the traditional/commercial model of software development and firm-based knowledge creation efforts, we show how the proposed model of knowledge creation expands beyond the boundary of the firm. Our model suggests that the product development process can be effectively organized as an evolutionary process of learning driven by criticism and error correction. We conclude by offering some theoretical implications of our community-based model of knowledge creation for the literature of organizational learning, community life, and the uses of knowledge in society. Revision to #OS 00-1246RR I.

Abstract not found

The Linux kernel development project was among the first attempts that make a deliberate effort to use globally connected software developers as the main source of talent and input to create an important, Open Source software. Based on the Linux project, we have built a model of Open Source knowledge creation to study how thousands of talented volunteers, who are dispersed across organizational and geographical boundaries, collaborate via the Internet to produce a knowledge- intensive product of high quality. Comparing and contrasting the Linux model with the traditional/commercial model of software development, we focus on four issues critical to software development: intellectual property licensing, incentive to contribute, coordination mechanisms, and production process. Recognizing that the applicability of the model may be constrained if business firms do not have the supporting infrastructure and work practices, we propose several areas where the model can be adapted and how the adapted models are useful to business firms. 2 I.

rare. We seek to draw concepts from open source for small systems engineering, and vice versa. But the bazaar methodology, in essence, exudes methodology. Although to some extent that's its point, this seems too chaotic and immature, even for small projects. A magazine article [3] expounding the virtues of the "grownup" methodology employed for the extraordinarily bug-free NASA space shuttle software described the dominant image of the software development world as: Gen-Xers sporting T-shirts and distracted looks, squeezing too much heroic code writing into too little time; rollerblades and mountain bikes tucked in corners; pizza boxes and Starbucks cups discarded in conference rooms; dueling tunes from Smashing Pumpkins, Alans Morrisette and the Fugees. Its the world made famous, romantic, even inevitable by stories out of Sun Microsystems, Microsoft, and Netscape. This kind of atmosphere is obviously not going to produce well-engineered software systems, outside a few

this paper, we assume that open-source generates a software product of same quality to that provided by proprietary development. One of the questions initially raised still holds, however. Namely, whether a software company can survive in the face of competitition from comparable open-source software, Current Address: Alphatech, Inc., Burlington MA 01803 which literally gives its product away. Are such companies doomed from the outset? Facing a "new" opensource competitor, would a company be advised to exit the market to avoid losses? The economic study of goods for which there is no charge (i.e. free goods) has been long studied. Hardin's (1968) described allocation of resources in the context of population growth and the use of natural resources (drawing upon studies dating back to Adam Smith). He described the "Tragedy of the Commons" in which an individual enjoys the full benefit of a free resource, but shares the costs with the rest of society. In this case, there is an incentive to overuse that resource (the Commons) until its ruin. Software markets, however, are unique in the sense that the marginal costs (i.e. the cost associated with an extra unit) are essentially negligible. Ghosh (1998) describes the trade of free goods on the internet in terms of a "cooking pot." That is, contributors donate their software to a common pot, the mixture of which may be freely sampled by anyone. Raymond (1998) proposes that the driving factor encouraging software donations to the internet pot is "noo," a combination of reputation and ego arising from creating and distributing a software solution to a specific need

Our Workgroup Cache system operates as a virtual intranet, introducing a shared cache to members of the same workgroup. Users may be members of multiple workgroups at the same time. Criteria are associated with each workgroup to pull documents from an individual cache to the shared cache, or push from the shared cache to an individual cache. These criteria provide semantics of the workgroup’s tasks and interests to reduce latency for its members. 1.