Two Bits: The Cultural Significance of Free Software (review)
In lieu of an abstract, here is a brief excerpt of the content:

Reviewed by
Two Bits: The Cultural Significance of Free Software. By Christopher M. Kelty. Durham, N.C.: Duke University Press, 2008. Pp. xvi+378. $84.95/$23.95.

Two Bits is an interdisciplinary hybrid: part theory, part ethnography, part history. It was not written for historians of technology, but it constructs an engaging history of free software that will appeal to readers interested in the [End Page 964] history of computing and, more broadly, in the collaborative aspects of technological systems.

I was relieved that Two Bits departs from familiar accounts of the economic logic (or lack thereof) in free software. Instead, its author, anthropologist Christopher Kelty, follows the words and actions of free software devotees who are committed to making source code publicly available and modifiable. Free software, in this cultural analysis, is not a stable product; instead it is a collection of practices. Kelty theorizes these practices as essential components of a “recursive public,” a term he has coined to refer to a group “concerned with the ability to build, control, modify, and maintain the infrastructure that allows [it] to come into being in the first place” (p. 7). Kelty defines a recursive public as “a collective independent of other forms of constituted power” (p. 3) that responds to power by producing working alternatives such as free software.

I appreciate Kelty’s insistence that free software is a “kind of collective technical experimental system” that reconfigures knowledge and power (p. 15), but I remain unconvinced that any recursive public could be truly independent of existing power structures. After all, free software runs on expensive hardware.

Two Bits begins with a cogent discussion of theorists on the fringes of the SHOT canon, including Jürgen Habermas, Charles Taylor, and Michael Warner. Kelty’s engagement with their work will be refreshing for historians of technology who crave more theoretical diversity. His “recursive publics” concept may find warm reception among his intended audiences (anthropologists and scholars of online communities), but I was frustrated by his explicit warning that he did not craft the concept to be more generally applicable (p. 305). Nevertheless, Two Bits highlights aspects of free software that might be interesting to compare to the practices of earlier generations of infrastructure workers, such as railroad engineers, telegraph and telephone engineers, radio hams, or the IBM SHARE user group. Perhaps concepts that Kelty does not discuss—Greg Downey’s “protocol labor” and Etienne Wenger’s “communities of practice” come to mind—will prove more useful for historians who seek to situate free software practices within a longer historical context.

Two Bits has some wonderful passages, including a brilliant chapter on “geeks” and the stories they tell about the Protestant Reformation. Specialists in the history of computing will be most interested in five chapters that sketch histories of the UNIX operating system, the ill-fated Open Systems Interconnection (OSI) network architecture, the Internet standards process, the GNU General Public License, and governance structures for free software such as Apache and Linux. Kelty relies both on familiar secondary sources and on a vast range of “born digital” primary sources preserved on the Internet. His assessment of his primary sources is exuberant: “Nearly everything is [End Page 965] archived. Discussions, fights, collaborations, talks, papers, software, articles, news stories, history, old software, old software manuals, reminiscences, notes, and drawings—it is all saved by someone, somewhere, and, more important, often made instantly available by those who collect it” (p. 21).

Seduced by abundance, Kelty fails to reflect on the ways that such “archival hubris” (p. 22) omits important offline moments such as telephone conversations, private email exchanges, and the ritualistic meetings in the hotel bar. Perhaps it is no coincidence that we have multiple histories of Linux and TCP/IP, which have a preponderance of source material at Google Groups and in the Internet Archive; and yet we lack detailed histories of Cyclades and OSI, which do not. The implication—and current historiographical reality—is troubling: those who shout loudest figure most prominently in histories of the information age.

I liked this book more and more every time I picked it up, and I recommend it along with Steve Weber...