Reproduced from the Journal of Peer Production by Christopher Kelty:

“Free software does not exist. This is sad for me, since I wrote a whole book about it. But it was also a point I tried to make in my book. Free software—and its doppelganger open source—is constantly becoming. Its existence is not one of stability, permanence, or persistence through time, and this is part of its power. Free software is valued for its peculiar form of potentiality. It is not any particular thing or technology or license: it is a possibility, a concrete utopia perhaps (Broca 2012). Free software promises a sequence of other values: experimentalism and creativity, provisionality and modifiability, rectification and refraction, dissent and critique, participation and obligation. Free software is not just process over product though—it is principle made material.

Indeed, I would argue that the mode of existence of any live technology (as opposed to those dead ones no longer maintained) is not singular or stable. It depends on how people and practices exist, how organizations and laws exist, how ideologies and discourses exist; and it is in constant motion. It is not “free software” that has critical power but something else—something that is lived and experienced, but which, to be honest, does not yet have a concept appropriate to it. I proposed the concept of “recursive publics” to make sense of it (Kelty 2008). It was a concept of limited utility, no doubt. It may or may not have captured the “critical power” of free software, but it was an attempt to articulate that power without hypostasizing or fetishizing free software (or open source) as such; and in order to give more critical specificity to analyses of contemporary capitalism than simply “information technology” or “new media” or “immaterial labor”—labels, not concepts.

The power of free software, however, is that it allows values and principles to be turned into material objects, things that can be manipulated, reconfigured, tested and torqued. In this it is no different than a legal reform, the formation of an institution, or the creation of an organization. But free software seems (or perhaps seemed) to have a peculiar power to leave this materialism radically open to change—material, but moldable, rigid but reconstructable. It turns the negotiation over values into a sub-field of maintenance, system administration, development, or design. It gives making an immediate, ethically inflected, political urgency. Furthermore it hails subjects who desire to engage (and take pleasure) in this kind of making: contemporary subjects of capitalism who want individual agency to be combined with a convivial but competitive collective experience (Coleman 2013). Free software is not a practice or a thing that seeks permanent consensus—but one that abhors stacking the deck, tilting the playing field, or the chicanery involved in manipulating law, technology and governance to the advantage of some group or other. All design, all maintenance, all administration is politics—this we’ve known for decades, if not always—but it has always been a sub-politics, a field of tactics, not strategy (Marres & Lezaun 2011). Free software has felt to some like strategy—like infrastructural and material strategy, like strategy for a new world.

The editors of this issue of the Journal of Peer Production have wisely suggested that the issue might be one of epistemology—that the power of free software, its concept (still unarticulated) concerns the epistemology of software development. But this is at once too narrow and too broad. It is too narrow because what advocates, adherents and proponents of free software see in it is something more than software—they see a style of remaking the world, and they immediately want to apply it to every aspect of life. They see something like Arendt’s “action” in The Human Condition—the chancy making of new beginnings, not the rote instrumentalization of politics (Arendt 1958). The epistemology of free software is the epistemology of liberalism updated—not neo-liberalism, but a liberalism of the 21st century, questioning institutions (like intellectual property) and configurations of power (like the financial sector or the Hollywood-Silicon Valley-Advertising complex) through reform and reconstruction. For better or worse it is the kind of liberalism that Jeremy Bentham and John Dewey—radicals both—sincerely believed in and sought to bring into existence.

But at the same time the “epistemology of software development” is too broad because free software was, at its height, the apotheosis of a very specific style and ethic of software creation central to the discipline of academic computer science. This certainly does not mean that it was the only one, nor that it had any necessary affinity with the dominant practices of software development in corporations, organizations or consultancies at any point between the tar pits of the IBM 360 and the agile development of the last decade (Brooks 1995; Beck et al. 2001; Hakken et al. 2010). Rather as I tried to demonstrate in my own work, UNIX culture came from a very specific and very unusual place—a hybrid academic-corporate culture within Bell Labs cross-fertilized by nascent computer science departments around the world, and sedimented in the minds of every budding computer scientist and software engineer by Andrew Tanenbaum’s textbooks. There were most certainly other traditions—but what is surprising is that they did not produce free software or anything much like it (but see Akera 2001). Free software—and UNIX culture before that—was always critical of those other traditions. I (and others) cannot seem to say it strongly nor repeat it often enough: that it is a culture and not just a development methodology. If free software (and UNIX before it) did not communicate values and a worldview, or if it fails to do so in the future, then it would almost by definition not have the power people have long associated with it.

But to repeat: free software is changing. The hybrid form that free software took—like all hybrids—can either survive or become sterile. Perhaps like the mule, free software is the end of the line; or perhaps what we saw in free software was a certain kind of hybrid vigor—a powerful expression of the cross-fertilization of the best of academic values and corporate efficiency, destined to live only once; or until another and similar mating occurs.
People who write, support, use or love free software today don’t want to hear this—but it should not frighten anyone. The reasons why people become obsessed with free software and its power come to free software, not from it. People recognized in free software a solution to a problem—to what I called all too vaguely “a reorientation of knowledge and power”—a disruption of the ways knowledge and power are related; a disruption in the creation, circulation, distribution and control of knowledge and how those things are remaking the landscape of power in which we all—hackers, corporations, academics and users—live and operate. Or to put it back in the crypto-francophone language it was originally derived from: free software is an assemblage, constructed in response to a problematization of knowledge and power, and dependent upon the dispositif of the knowledge-content-technology industries, the intellectual property system, and the educational institutions that feed them both (Foucault 2009, 2009, 2012; S J Collier 2009; Rabinow 2003; Ong & Collier 2004). Each of these elements change at different rates, and capture a different scale of the cause and effect—with free software being the most local and unstable of them. 1998 was thus an inflection point at which an assemblage formed—like a precipitant or an emulsion—out of the component practices that were all part of the dispositif, but struggling to deal with a changed set of conditions—anti-IP legal activism, collaborative software development, liberal ideologies of competition and capitalism, and engineering conceptions of open, interoperable infrastructures, among other things. A question then might be posed: is free software an assemblage of its own, or is it one component in a larger assemblage? Is it just one tactic among others, alongside the pranksterism of Anonymous, the good fight of EFF, the reformers of copyright law, and the global movement for an open, re-writable, re-mixable culture? Or is it—along with open source—a strategy, a core or underlying practice that we don’t actually value or understand deeply enough, but without which all the rest is sound and fury?

In 1998, “open source” was indistinguishable from free software—it was literally nothing but a different name for the same thing. That is no longer true in 2013. But “domesticated” open source is not a decimation of the power of free software, it is a different solution to the same problem. Perhaps it is a solution benefitting other and now richer and more powerful actors. In the 1980s and 1990s corporations too were fighting over the relationship of knowledge and power. Within the ecology of legal forms of intellectual property they inhabited, they brokered a kind of truce in which every corporation was forced to re-invent the wheel, at immense cost, in every production cycle. Open Source solved this problem—and for a great many software developers, toiling as they do in the richer veins of freelance precarity, it meant not having to rebuild the same damn thing over and over again with every upward career move. And so it was better for them and for the corporations—even if neither of them saw the radical potentiality of free software, but only the humdrum advantages of open source.

The link, therefore, between a powerful software-based practice of techno-infrastructural reform, and a public-oriented, critical and politicized one grows more tenuous with every free software developer enticed by Apple, Google, Amazon or Facebook by the promise of a stable work environment and a regular salary. They join not because they have abandoned their ideals, but because, perhaps, they cannot abandon their stomachs or their children. Does this mean free software (contra open source) then becomes more and more the property of radicals and activists (who equally have stomachs and children), breaking that link that was once so central to the practice?

If the critical power of free software is now more strongly, and globally, associated with hacktivism, hackerspaces, Tor, Anonymous—and more with Europe, Asia and Brazil than the US—it is because it has concentrated on a particular ideological goal: the copyfight. In terms of what Coleman and Golub called “genres” of hacking, the meaning and focus of free software has become more and more the domain of critiques of intellectual property, and in particular critiques in the name of free speech, privacy, security and the individual values of liberalism. Add to that the occasional hopes of Occupy and its ilk to transform infrastructures or create global commons or decolonize our technosphere (Coleman & Golub 2008; Schoonmaker 2007).

By contrast the “critical power” (if such it is) of open source operates in a different genre—one associated with problems of infrastructure, corporate continuity, competition, maintenance and support (Jackson et al. 2011; Jackson et al. 2012). In the 1980s and 1990s, the Internet and free software were scarcely distinguishable. There were microcomputers and PCs and modems and routers all connected by the same networking protocol. The reasons why that protocol spread explosively in advance of other hoary, established, international standards were the same reasons that Unix, Minix and Linux spread around the world: the potentiality, modifiability, and portability of the tools and code. It was a world of university engineers and corporate subversives spouting rhetoric—rhetoric also internal to the commercial industry—about the malady of giant corporations and the rapacious capitalism they were practicing.
Back then, however, free software was never anti-capitalist; even in its strongest, most ideological forms it required assent to, and belief in, the power of the ideas of classical liberalism, especially that of free expression. It did this in order to save capitalism from the capitalists—to point out its failings and reintroduce certain kinds of limits in the name of freedom and equality. But today, free software is more often associated with revolutions, with the movements for global justice, and especially by some academics, with critiques of capitalism as such—a tendency that was recognized early on (e.g. in the work of the Oekonux group) but which has only recently started to become more widespread.

That a split between free software and open source has occurred is surprising, but predictable. The legal and technical environment we live in has changed—the dispositif is slowly changing, and so the elements that made up the assemblage of free software are also shifting—separating open source and free software in new ways. For the first time there is a concrete material difference, and not just an ideological difference, between the practices of free software and those of open source. It is enough of a difference that the label FOSS or F/LOSS no longer works. The open source “genre”—that of infrastructure, standards and competitive commercial power—is still vibrant today. Web frameworks like Ruby on Rails; virtualization tools like Hadoop, mobile platforms like Android, and even something like Facebook Connect represent the “critical power” of “open source” today. They are all built (of course) on the solid foundations of Linux servers and open source networking tools, email services, programming languages and editors, tirelessly maintained in common, so that they don’t have to be re-invented again. But Free Software advocates scarcely recognize themselves in the success of Facebook or Google, the privacy policies they implement, or the engineering strategies they pursue.

A key reason is that the basic structure of the Internet has drifted far from the ideals of its origins—it is now a small, tight web of powerful ISPs and Telecommunication giants and data centers, not a distributed network (despite the ever-present popularity of mesh networking, peer to peer algorithms, and DIY, low-power, backyard creativity); its giant backbones are guarded like oil pipelines; its basic structure can be manipulated at the national level (as we saw recently with Egypt and Syria and routinely observe with China and Saudi Arabia). We have proliferated technologies that are aimed at shoring up security holes and providing user convenience at the expense of other principles.

Even more significant is the transition out of the era of the desktop PC—into an era heralded for as long, indeed even longer, than the PC has existed: computing as a utility, virtual computing, network computing, thin clients, MetaComputing, Grid Computing, application service providers, web services, service oriented architectures, software as a service, cloud computing, and so on. The idea of remotely accessed software services on the model of a pubic utility has been around since Douglas Parkhill’s 1966 The Challenge of the Computer Utility. Time-sharing was a reality by 1965; IBM invented sophisticated virtualization (CPS/CM) for its famous OS/360 in 1968; remote procedure calls were part of early ArpaNet RFCs and a core feature of the SUN workstation by 1985; CORBA was a much-discussed, and abhorred, standard in the 1990s and web services and http-based protocols for web-based computing (like SOAP and REST) have dominated the learning curves of programmers for the last 10 years.

The current success of “the cloud,” however, is more specifically about the material transformation of the Internet from a heterogeneous ecology of small and medium sized nodes (single servers and small clusters of servers) to a handful of enormous data centers and server farms running virtualized versions of the Internet inside them. The biggest of these, like Google, Amazon or Rackspace are all running “open source” software—but the concept makes less and less sense the larger and more controlled these systems become. If the critical power of open source was about its openness, its potentiality, or its modifiability, then the counter-critical power is simply monopoly; control all the servers and even if they all run open source software, they only do so at the pleasure of the sovereign, so to speak—the once beloved power to “fork” the software disappears in a world where centralized infrastructure is everything.

In the consumer space, a similar change has occurred: what difference does it make if iTunes or Android Marketplace runs on free software? Nobody really liked the desktop version of free software anyways, right? So now we must submit to Apple’s dominion to play (or Google’s or Amazon’s). A few of the most committed free software advocates weep a little every time they think of this state of affairs, but the 99% increasingly find no personal or political problem with this state of affairs—and perhaps they are right not to, perhaps the issue is all too academic. Philosophers have a story for the kind of freedom at stake here: the problem of the contented slave (Pettit 1997; Pettit 2001). The point, however, is that free software can no longer be a recursive public in these spaces—that the fragile, critical power it had evaporates from the heat of massively centralized data centers; it is repulsed by the sheer gravitational mass of the monopolies forming anew; it is dissolved in the frantic and frustrating need for our devices to function from one minute to the next—as in Arendt’s story, the need for maintenance (labor) overwhelms the need for freedom (action).

These evolving genres of free software and open source have also been taken up globally. The promise of open source as a solution to problems of infrastructure and standardization has functioned as a doorway into the industry for those historically locked out. Rather than a national competition, with national standards—the 1980s and open source ushered in, with “flexible capitalism” generally, a global labor market in which software developers in southeastern Brazil, Bangalore, Malaysia and eastern Europe started mixing and collaborating more often than being body-shopped or off-shored (Xiang 2007; Murillo 2007; Murillo 2010; Takhteyev 2012; Chan 2007; Chan 2009; Saxenian 2007). For better or worse, open source—as a response to the reorientation of power and knowledge—benefitted new classes of global citizens at least as much as their Euro-American compatriots and brought them into new modes of affinity, collaboration and competition. But as open source is routinized, so too are the lives of those newly inducted; as open source becomes an instrumentalized kind of politics, the possibility of new beginnings fades.

Meanwhile, the promise of Free Software has similarly seen global uptake: as part of the Global Justice movement, as part of the rise of Anonymous, Wikileaks and LulzSec (Coleman 2012); as part of the counter-cultural spread of hackerspaces, media labs and DIY/maker culture around the world (Lindtner & Li 2012). The distillation of the political power of Free Software into these radical, alternative and counter-cultural spaces and communities, however, means that it is further and further de-hybridized—no longer so solid a land bridge between the corporate forms of intellectual property-saturated IT industries and the cultural uptake of software and tools; no longer a wrench in the works, perhaps, but merely a wrench in the upraised fists of protesters.

But the changes facing free software are neither incessant nor incomprehensible. The phenomenon is too widespread, too entrenched, too familiar and too powerful to simply disappear. What we lack—scholars, activists, developers, lawyers alike—are concepts appropriate to this phenomenon. I insist here on discussing the differences between open source and free software because I worry that scholars—especially scholars who see themselves as radicals or liberals—are content to study and support only the increasingly radical, underground oppositional forms that free software takes—and to avoid a careful cultural analysis of the domesticated forms that open source is taking. Contemporary critiques of immaterial capitalism or communicative capitalism or whatever the latest branding is, fail miserably at the task Marx once set himself: to describe the workings of capitalism in such excruciating detail that the critique could not help but emerge from that description. We need an analysis that gives us the concepts with which to understand what this industry and its infrastructures is doing to the world; as it centralizes and re-monopolizes for the n-th time in as many decades; as it becomes ever less distinct from the advertising world; as it converts well-spoken, critical programmers and sys-admins to silenced employees; as it becomes a form of welfare for global populations of precarious high tech workers; as it becomes a more and more extractive industry—extracting time and money in-game and out; as it generates within itself the very tools for its analysis, transformation and reconstruction.
There is no free software. And the problem it solved is yet with us.”