"The Race for Machine Intelligence: DARPA, DoD, and the Strategic Computing Initiative"
by Andy McIntire
Alex Roland from Duke University started Carnegie Mellon's Spring 2000 Cold War Science and Technology Colloquium in a presentation co-sponsored by the Department of Engineering and Public Policy and by the International Software Research Institute.
The talk was a personal tour through a two year research project in which Roland and Philip Shiman wrote the history of the Strategic Computing Initiative (SCI). The results of that research will be published as Only Connect: DARPA, Strategic Computing, and the Integration of High Technology, 1983-1993, probably out later this year.
The Federal government, through the Defense Advanced Research Projects Agency (DARPA) spent $1 billion dollars from 1984 through 1993 in an effort to spur the development of intelligent computers, ostensibly as key components in automated battlefield technologies, though frequently the final design goals were somewhat contrived as a means to sell the project. Professor Roland recounted the basic "vanilla" history of the project as a prelude to a discussion of the interpretive paradigm problem he and Shiman encountered as they progressed.
First, the story of SCI can be told as a piece of the larger picture of American technological history. This approach sees SCI as a reaction to the post-Vietnam syndrome, Ronald Reagan's efforts to ramp up the Cold War after his election in 1980, as an American effort to counter Japan's much-publicized 5th Generation computer initiative, and in short as a logical reaction to contextual events. Another framework the authors considered was a social constructivist perspective centering on an ethnohistory of DARPA as a culture within the computer science community, and considering issues related to how the organization defined objectives and sought to proceed toward them. Alternatively, the story could be told, Professor Roland explained, as a straightforward institutional history, or even as an example of the Military-Industrial-Academic complex at work.
Still yet, the story could be told so as to foreground issues related to research patterns, asking questions about what kinds of work government supports, and with what results. Another alternative would be to see the tale as one of the contrasting management styles of the major players involved, and their conceptions of research. Or, the story could be a case study in National Industrial Policy, returning to the old debate over government actions in the marketplace.
Alternatively, the SCI story can be seen as a case study in Dual Use Technology: military research is often seen as the source of spinoffs, and debate over the pursuit of military and/or civilian needs is quite relevant. SCI also illustrates the Metrics of Patronage because rich societies usually invest in the arts and sciences, and in Cold War America the Federal government acted as patron for much scientific research.
In the end, however, Roland and Shiman told the story of the Strategic Computing Initiative as one of technological trajectories. Professor Roland explained that the story of SCI was one of a governmental attempt to identify the course of computer technological development so as to intervene to speed development. This paradigm incorporates and subsumes many of the other paradigms outlined above; tracing the technological trajectory prior to SCI required the authors to deal with context, for example, and helped them explain the social construction of the project.
Alex Roland met with the Cold War Group for lunch before his talk. Clockwise around the table are: Alex Roland (at head of table), David Hounshell, Gerry Fitzgerald, Glen Asner, Asif Siddiqi, Ruud Van Dijk, Joel Tarr (closest to camera), Doug Davis, Kiron Skinner, Ed Constant, and Jim Tomayko.
[Text Only Index]