Weber’s landmark book analyzes the open source movement from the perspective of political economy. Tracing the history of the movement from the Unix hacker culture of the 1970s through the birth of the legendary open source initiatives of the 1990s to the present (he’s writing in 2004), he is primarily interested in analyzing the factors of the open source process that have lead to its success in producing high quality, extremely complex products without any hierarchical structure and very little monetary incentive. This presents a significant challenge to some of the basic precepts of political economy, according to which such success should be impossible: For one thing, the making of nonexcludable goods (those for which no one can control access) and nonrival goods (those which can be copied indefinitely without incurring expense) is normally disincentivized in a capitalistic economy; for another, the commons is generally thought to tend toward underprovisio because it admits of many more freeriders who deplete it than producers who contribute to it; and finally, hierarchical structure is customarily thought to be more efficient at producing large-scale products of significant complexity than distributed structures. Open source appears to turn such conventional wisdom on its head: in fact many thousands of programmers are motivated to work on such projects even though the vast majority receive no monetary payment for doing so and must balance such work with other employment, the open source commons seems to grow more robust the more freeriders it attracts, and distributed structures have produced arguably the most complex piece of software ever written: the Linux operating system kernel. Additionally, open source seems to overcome another piece of conventional wisdom specific to the software industry, codified as “Brooks’ Law”: Projects become more chaotic and further behind schedule the more developers you add to them. That is, productivity doesn’t scale at the same rate as vulnerability to bugs. Open source projects seems to have found a way to circumvent this “law”!
Weber maintains throughout his book that open source is a process, not a product. It can and has produced some very impressive software, but it could also produce many other products. Thus it is vital that we understand how it works, and in what ways it is exportable to other domains of activity. Weber’s analysis takes up two main strands: the incentives of the individual participants (why do they do it?) and the structure of the organizations (why is it so efficient?). To answer both, it is important to understand open source’s relationship with and stance toward intellectual property (IP). Some open source hackers, notably Richard Stallman and the members of the Free Software Foundation, are staunchly opposed to intellectual property law, believing that it is unfair and prevents real creative work from taking place. Hence the FSF’s opposition to all non-free software. Most open source developers, however, are pragmatists who are fine using (and sometimes even developing) closed source software as long as it is high quality. What is paramount to open source developers, claims Weber, is the quality of the code. Most believe that open source as a process produces better code than closed-source. Why? For one thing, it is empowering. Because no one “owns” the code, no one can be excluded from it. This means both that anyone can take the code and do whatever she wants with it, and that her work will not turn out to be for nothing because a company goes out of business, makes an executive decision to stop developing that particular piece of software, etc. Open source programmers do their work voluntarily, working on whatever part of the code they wish to work on, and they can be confident that their work will reach other users. Plus, other programmers and users will see not only the result of this work (in the form of programs that run) but the work itself in the form of code. Not only does this encourage higher quality work, but provides a platform by which a programmer’s skills can be seen publicly: “The programmer participates in an open source project as a demonstrative act to show the quality of her work. Reputation within a well-informed and self-critical community becomes the most efficient proxy measure for that quality.” (142) While IP generally functions by making certain intellectual goods excludable (and is therefore meant to incentivize the creation of such goods, as they can then be monetized), open source incentivizes developers to give away their work in exchange for recognition of the quality of that work. The primary legal mechanism for ensuring such openness is the software license. While some licenses, such as the BSD license, merely allow any work covered by them to be used however anyone wishes to use them (including integrating them into closed source products), other licenses, such as the GPL, actually prevent anyone from using the code for any project that is not itself open source. Here, copyright law is utilized to ensure that no one can ever own derivatives of the code produced, and property is re-conceived as a right to distribute, not a right to exclude.
The result is that open source software becomes a nonexcludable, nonrival good, a category that has its own peculiar characteristics. One that Weber emphasizes is its positive network externalities: as the software is used by more people, its value increases to its developers. This is why the commons created by open source isn’t depleted by freeriders or underprovisioned: in fact, “the system as a whole positively benefits from free riders.” (154) Thus Weber refers to open source software as “antirival”: it encourages freeriding. Linked to this is the more prosaic observation that open source blurs the distinction between producer and consumer. Though not all users of open source software are producers, the producers are themselves users, and in fact one of the most frequently cited motivations for starting an open source product is to solve a problem that the developer faces in her everyday life. That is, one of the primary motivations for open source development is “to scratch an itch.”
Weber compares the open source process to the process of evolution: both proceed amidst seeming chaos and produce highly refined products through a dual process of variation and selection. Open source is very good at both processes because of the diverse and distributed nature of its participants, the fact that they are both users and producers, and the fact that any piece of code in the digital commons can be utilized for any other piece of software: open source, then, is radically modular and recombinatory. Here Weber invokes Conway’s Law (“the structure of the [technical] system mirrors the structure of the organization that developed it”), noting that the macro social structure of open source projects is modularized in the same way as most of its software products (notably Linux and Apache). (174) These, then, are both engineering and social values: simplicity and modularity. They are, in fact, essential values for large-scale, complex projects utilizing distributed innovation. Each developer or team works on their module, focusing in particular on its inputs and outputs: it must be able to accept X information and produce Y information in order to interface with all of the other modules. Not only does this allow diffuse organization and distributed innovation to produce technically complex products by building from the bottom up, but also encourages innovation in each module because the system is not monolithically interdependent: it is more difficult to produce a cascading bug that destabilizes the entire edifice, and thus developers have some assurance that they can innovate without fear. In hierarchically-organized projects (such as most closed-source projects), the product is itself interdependent (Weber notes that the more complex projects become and the more time pressures come into play, the more interdependent and less modular software projects tend to become; open source counters this tendency through its organizational structure) and its developers are often afraid of radical innovation because there is too much to lose by introducing an unexpected bug into the software.
The primary advantages of open source, then, are related to its evolutionary potential, and sharply contrasted to top-down engineering. Open source proceeds with what seems like chaos, but consistently produces good code because it is transparent and modular: anyone can check anyone else’s code, find bugs, modify it, fix it, or use it for something else. Open source creates an ecology of code as well as a social structure that incentivizes participation. The result is rapid innovation: code is tested, modified, recombined, “mutated,” and speciated (forked) at an extremely high rate, and the result is innovation at many levels. By contrast, top-down engineered code, created within hierarchical structures, doesn’t participate in the same sort of code ecology: far fewer people can see the source, and therefore it becomes a secretive, conservative process. Weber analyzes this in relation to networks: some are intelligent, making centralized decisions about how information is routed, and some are stupid, or neutral, and merely deliver information without any active routing. The telephone network is an example of the former, the Internet the latter. Closed source is centralized: decisions rest with a few people only and information is routed according to their dictates. In open source, all bits of information reach all others: no intelligent routing happens on the Internet. The intelligence, instead, moves to the edges. The end user is empowered in this sort of network, and when the end user is also a producer (as in open source), the result is “distributed innovation.” This is not merely division of labor. Innovation can happen at any one of the nodes independently: “There are no weak links in this chain because there is, in a real sense, no chain. Innovation is incentivized and emerges at the edges; it enters the network independently; and it gets incorporated into more complex systems when and if it improves the performance of the whole.” (233)
Weber believes that this process is exportable to other domains of activity. He provides fictional scenarios in the petrochemical industry, medical profession, and genomics (distributed processing of genetic code). He believes that open source can provide an enormous advantage wherever contributions can be derived from open, accessible information, where the product is perceived as valuable to a critical mass of users and can improve from their attention and review, where there are strong positive network effects to use of the product, and where a voluntary community can develop around the process of building it. The greatest change that open source entails, however, is in its notions of property. That is, this process depends entirely upon a mindshift: away from property as right to exclude, and toward property as right to distribute. This is a radical shift that is difficult for many individuals and firms to conceptualize, and is the key enabler of the process from both a technical and social point of view.
Purchase the book here (Amazon).Tagged with: BSD • commons • copyleft • distributed innovation • evolution • GPL • innovation • nonexcludable • nonrival • open-source