Michael D. Gordin reviews a history of the Soviets' failed national computer network.
The 'small-n' challenge plagues all historians, but this problem of under-sized samples is especially acute in the history of science and technology. Most scientific discoveries seem to happen uniquely. We do sometimes see multiples — about half a dozen articulations of the principle of the conservation of energy or the periodic table around the same time, for example — but the diversity of specializations, the pace of communication and the vagaries of publishing mean that most innovations arise as singletons, to use sociologist Robert Merton's term. The issue is perhaps most striking with the flagship technology of our present moment: the Internet. Here, we have an n of 1.
This matters for two reasons. First, singletons frustrate generalization, making it difficult to draw lessons for science policy. Second is the related puzzle of contingency. We currently have an Internet, and it has a set of properties (such as the end-to-end principle, which stipulates that applications should happen at the edges of the network, rather than at intermediary nodes). Does it look like this because it has to, or are its features contingent characteristics of this specific Internet? Without alternatives to compare it to — a larger n — we just cannot say.
Soviet computer scientist Victor Glushkov pushed to create an 'all-union' system.
In How Not to Network a Nation, communications specialist Benjamin Peters argues for contingency, on the basis of an n increased from 1 to 2. Well, to 1.37 or thereabouts.
Historians have already started to chronicle networks past as useful comparisons. One is Project Cybersyn, an experiment to network the Chilean economy under president Salvador Allende in the 1970s, described in Eden Medina's Cybernetic Revolutionaries (MIT Press, 2011). Peters summarizes these well, but his quarry is the great white whale of this specialized historiography: the Soviet Internet. Whether there ever was such a thing, why it never moved beyond the project stage and which of the various projects between the late 1950s and the late 1980s can be properly classified as efforts to develop one are the main subjects of the book.
Peters makes a good case to move beyond historian Slava Gerovitch's excellent pun on this seeming oxymoron: “InterNyet”. His intuition is spot on. The cold-war origins of the US networking programme have been well documented, for example in Janet Abbate's Inventing the Internet (MIT Press, 1999). Direct military sponsorship was crucial. The defence department provided patronage through its Advanced Research Projects Agency, which launched ARPANET, the embryonic Internet, on 29 October 1969. The Internet's conceptual roots included cybernetics, created by mathematician Norbert Wiener in 1948. Given the close parallels between Soviet and US cold-war technologies, it would be surprising if there had been no efforts to generate a Soviet Internet. Indeed, Peters finds six different proposals to develop an 'all-union' computer network. This stands to reason, given what he calls “the outsized infrastructural imagination of Soviet planners”, who liked their projects big and utopian — think the space programme, dams and nuclear power.
Peters concentrates on computer scientist Viktor Glushkov's OGAS (obshche-gosudarstvennaia avtomatizirovannaia sistema, as the standard Library of Congress transliteration would render it — although various systems are used inconstantly throughout the book). The full name is a mouthful: “All-State Automated System for the Gathering and Processing of Information for the Accounting, Planning, and Governance of the National Economy, USSR”. Beginning in 1962, Glushkov spent 25 years trying to mobilize support for his network from his Institute of Cybernetics in Kiev, which created a rich set of cultural resources, including a model constitution, passport and cartoons depicting the land of “Cybertonia”. Peters reproduces these in plentiful images and descriptions, chronicling their utopian spirit and demonstrating the need for engineers in all times to let off steam through flights of fancy. But the project was never realized.
It is difficult to glean all the technical specifics from the material that Peters mobilizes from archives, interviews and declassified CIA reports. Some proposals look like cloud computing or tablets, but it would be anachronistic to interpret them in that way (and Peters doesn't). The idea was to use real-time processing to connect economic inputs and outputs, rendering the planned economy both functional and adaptive. We cannot even be sure that Glushkov's plans would have worked. What we do know is that the failure was not caused by a scarcity of personal computers, because OGAS was meant to link factory mainframes. Nor was it ideology: cybernetics, as Peters readably recounts, was well suited to Soviet ideological preferences in materialism and planning. To discover the roots of the issue, Peters invokes the cybernetic concept of heterarchy, which he defines as “complex networks with multiple conflicting regimes of evaluation in operation at the same time”; he then uses this to explore the heterogeneity of approaches to networking.
Perhaps predictably, OGAS's demise was death by a thousand paper cuts. Documents were misfiled, meetings were missed, the military and the statistical ministries disagreed about who would benefit. Peters's provocative thesis is that “The capitalists behaved like socialists while the socialists behaved like capitalists.” The US Internet was the result of state subsidies and benevolent paternalism; the Soviet attempt foundered on entrepreneurial infighting. (Elsewhere, Peters puts the culprit down as cost, although how costs were tabulated was in itself a bureaucratic conundrum.) There is no dramatic climax to How Not to Network a Nation. Non-existent technologies end with a whimper, but even whimpers can tell you something.