Learning from Linux Internet, innovation and the new economy Part I: Empirical and descriptive analysis of the open source model Working paper, April 15, 2000. Please do not quote without permission. Ilkka Tuomi [email protected] Abstract: This paper describes some key principles that underlie development of the Linux operating system, and discusses their implications for innovation research. Using Linux as a case example, I will discuss organizational, institutional, economic, cultural, and cognitive aspects of the open source development model and technological change. 2 Introduction According to user surveys, the Linux operating system is rated as the best operating system available. It is considered to be more reliable than its main competitors. Its functionality is claimed to be better, and according to many experts, new releases of Linux implement innovative ideas faster than its competitors. In other words, it is argued that Linux development creates complex new technology better and faster than the biggest firms in software industry.1 Yet, Linux also seems to break many conventional assumptions that underlie research on innovation and technological change. Linux is developed by an informal self-organizing social community. There is no well-defined market or hierarchy associated with it. Most of Linux development occurs without economic transactions. Instead of getting paid for their efforts, the developers often spend a lot of money and effort to be able to contribute to the advancement of the development project. The open source development model, which underlies Linux, has attracted increasing attention in the last years. Today, Linux is considered to be a serious threat to Microsoft’s market dominance in operating systems. More generally, open source development projects have in recent years had major impact in software and internet-based industries. For example, over 60 per cent of Internet connected Web servers were open source Apache servers in March 2000. As can be seen from Figure 1, the second most popular Microsoft servers were about one third as popular with 21 per cent. Some open source projects, such as Sendmail and Emacs, have achieved large market shares, making it difficult for commercial enterprises to enter the market. The Internet Engineering Task Force, which defines standards for the Internet, has also used an open source approach since its formation in 1986 (Bradner, 1999). Several commercial software firms have tried to adopt aspects of the open source model. For example, Netscape announced in 1998 that it would distribute the source code of Netscape communicator with open source 1 http://www.uk.linux.org/LxReport.html 3 license. IBM decided to use the open source Apache server as the core of its Web server offers. Red Hat, in turn, thrives on packaging Linux distributions, and producing added value for Linux users. In all these cases, business firms are experimenting with ways to benefit from innovation that occurs in the open source communities. Instead of traditional economic competition, such initiatives rely on symbiotic relationships, and on the willingness of developer communities to collaborate. Figure 1. WWW servers connected to the Internet.2 This paper focuses on the Linux operating system kernel. I will describe its history and development model to discuss the organization and drivers of innovation in modern global economy. As the paper is heavily based on one specific software development project, I will also discuss the possible ways the learnings from this specific case may be generalized. A subsequent paper will further continue this work by linking the empirical results of this paper to the theory of innovation, research on product development, and technology policy. In much of the innovation literature, innovation is defined as something that has economic impact. The case of Linux shows that this definition is problematic and possibly misleading in important practical cases. During its history, most Linux development has occurred independent of direct economic concerns. It would be tempting to argue that Linux development is different from “economic activity” and something that, strictly speaking, 2 Source: Netcraft, http://www.netcraft. com/survey/. 4 should not be called innovation. Indeed, Linux development has not in any obvious way been associated with changes in production functions, market competition, or appropriation of economic investment and surplus. This, in itself, makes Linux an interesting test case for economic theories of innovation and technology development. For example, the case of Linux allows one to question to what extent existing economic models of innovation and technological development capture phenomena that underlie collective production of new technologies. In very practical terms, Linux is an economically important phenomenon. Indirectly, the success of many new businesses, venture capitalists, investment funds, and individual investors critically depends on the productive activities of the Linux community. Yet, when we consider the entire history of Linux, the economic impact seems to appear almost as an afterthought and as a side effect of a long period of technology creation, in a way that seems to break commonly accepted rules of innovation and technology development. Linux, therefore, provides an interesting history of globally networked innovation, illustrating the substance that underlies the discussions on the “new economy.” From a theoretic point of view, Linux is an interesting case as it enables us to discuss those social and cognitive phenomena that underlie technological change. In that sense, it enables us to penetrate some black boxes of innovation theory, including such widely used concepts as learning, capability, utility, and consumption. By observing the development of Linux, we can describe the microstructure of innovation, and transcend the boundary between invention and innovation. Linux development is collective productive activity, and the products of this activity are externalized as technological artifacts and discourses, which can be observed relatively easily. There exists sufficient documentation on the history and practices of Linux development so that we can—at least tentatively—describe some key principles that underlie Linux development. From a practical point of view, the case of Linux also provides a test case for analyzing product development models and proposals for organizing for innovation. Specifically, the extensive use of modern communication and collaboration technologies in Linux 5 development highlights some aspects of technology development that were not easy to see in earlier studies on innovation. This paper is organized as follows. First, I will briefly describe the Linux system and its developer community in an evolutionary context, highlighting some main characteristics of the socio-technical change that has led to the current Linux system. I will then discuss the organization of this technology creation process, focusing on control and coordination mechanisms. I will describe in some detail the ways the Linux community has managed the trade-offs between innovation and maintainability of the increasingly complex system, and discuss how the learnings of effective coordination and control mechanisms have been embedded in the system architecture. Linux has attracted considerable attention because it has been argued that the open source quality control mechanisms are more effective than traditional methods used in software development. It has often been argued that Linux is more reliable than proprietary systems because it is developed using the open source principles. I will describe the quality control system, analyzing in some detail the Linux bug removal process and the complex socio- technical system that underlies it. As was noted above, innovation literature sometimes leaves the process of invention into a black box where undefined psychological forces operate outside the domain of innovation research. The drivers for innovation are commonly understood to be economic. In this context, it is interesting to analyze the incentives and drivers of technology development, as they can be observed in the case of Linux. I will do this, describing the reputation and attention management processes that underlie Linux. Reputation and attention are shown to be closely related in the Linux community, and they are the key to resource allocation, which, in turn, directs technology development. Open source development is a special form of technology development as it intentionally reverses some common intellectual property rights. Instead of copyright it uses “copyleft,” which guarantees the rights of users to modify the results of development, and derive new works from it. The fact that such a licensing model seems to work and promote 6 technology development has important consequences for discussions on intellectual property rights, patent system, and the theory of appropriation of the results of innovation. The open source licensing policy can be seen as an important social innovation that has major impact to the way Linux is developed. I will discuss regulations and standards that underlie Linux development, focusing on the various forms of licensing that have been used in the open source community. Finally, I will conclude by drawing together some results from this descriptive and empirical analysis, putting them into the context of the global dynamics of modern economy and innovation. The