Henry Petroski reviews 'The Department of Mad Scientists' by Michael Belfiore

By Henry Petroski
Sunday, January 3, 2010


How DARPA Is Remaking Our World, from the Internet to Artificial Limbs

By Michael Belfiore

Smithsonian. 295 pp. $26.99

After the Soviet Union's 1957 surprise launch of the world's first artificial satellite, Sputnik, the U.S. government responded by establishing the Defense Advanced Research Projects Agency. Sputnik showed the United States that it was lagging in space technology, and DARPA's mission was to prevent future technological surprises. It soon became a corollary objective of the agency to create surprises of its own -- as well as world-changing technologies. In this regard, DARPA's record has been stellar, with such achievements as early versions of the Internet and the Global Positioning System to its credit.

Anticipating and producing surprises necessarily meant a certain degree of secrecy, at least during early development stages, but DARPA now also directs and funds a good deal of research out in the open. Understandably, this more public work is the principal focus of Michael Belfiore's book, which he and his publisher remind us is the first "inside look" at the organization.

DARPA is an unusual government agency, in that it makes a deliberate effort to prevent its program managers from becoming an entrenched bureaucracy and minimizes the amount of red tape associated with its projects. This enables quick and responsive funding decisions and the kind of risk-taking that can indeed produce world-changing results.

In addition to providing insights into the nature of the agency itself, Belfiore focuses on several historical and ongoing DARPA development initiatives besides the Internet and GPS, including electroprostheses for Iraq veterans, remote robotic surgery, driverless trucks and hypersonic flight. Such efforts, which are driven by the needs of the military, ultimately can have civilian benefits.

Because so much of what DARPA funds has a clear and practical objective, it often involves more engineering than science. In this regard, the book's title is misleading. It is not mad scientists who are engaged in much of the work but rather sane engineers.

DARPA has had identity crises in its half-century of existence. The fall of the Soviet Union in 1989 raised questions about the continuing need for a defense-driven strategy, and so in 1993 the "D" was dropped to leave ARPA, which was actually the name it bore from its beginnings to 1972. The resulting less-focused agency was short-lived, and it returned to being DARPA in 1996. The organization's legacy has been emulated recently, when the Department of Energy established an Advanced Research Projects Agency-Energy (ARPA-E) to promote research in such areas as biofuels, batteries and solar power -- work that DARPA was already engaged in for military applications. Cooperation between DARPA and ARPA-E would obviously be in the best interests of both agencies and the nation, and there is precedent for undertaking such cooperative research and development efforts.

Among the more developed case studies in Belfiore's book is DARPA's involvement with the Internet, which started out as the Arpanet, a relatively small network of interconnected computers at universities and research laboratories engaged in DARPA-funded research. At the same time, there was increasing interest in making computer-user interfaces friendlier and in allowing remote time-sharing of large mainframe machines. The foundations for these goals and more were set down in the 1960s through cooperative funding from DARPA, NASA and the Air Force Office of Scientific Research.

In a historic occasion at the Fall Joint Computer Conference in 1968, Doug Engelbart of the Stanford Research Institute demonstrated before an audience of thousands of computer researchers the rudiments of "word processing, video chat . . . real-time collaborative document editing, hypertext links . . . and the computer mouse." The Arpanet was mentioned as a new project at that event, too.

Another extended case study relates to DARPA's Urban Challenge, which involves computer-controlled supply vehicles negotiating streets, traffic and obstacles of the kind that can be expected in a war zone. The challenge takes place in the Mojave Desert, where a mock urban area has been created. University and other teams compete against each other in the challenge, whose military application is clear. In addition, the spin-off benefits of terrain mapping and collision avoidance technologies can be expected, like so many of DARPA's efforts, ultimately to yield considerable civilian applications.

Henry Petroski is a professor of civil engineering and of history at Duke University. His new book, "The Essential Engineer: Why Science Alone Will Not Solve Our Global Problems," will be published in February.

© 2010 The Washington Post Company