Opinion

• Book information
• The Department of Mad Scientists by Michael Belfiore
• Published by: HarperCollins
• Price: $26.99

ON 6 December 1957 a hollow aluminium sphere the size of a small melon burst from a blazing fireball, rose a mere metre or so above Florida before landing with a thump. The US was in trouble. A month earlier, the Soviet Union had sent a 500-kilogram capsule bearing a dog called Laika into space. But here was the US unable to even notch up its first foray into orbit.

President Dwight Eisenhower responded by creating a new research agency tasked with ensuring such "technological surprises" like Sputnik would never be sprung on the US again. The Defense Advanced Research Projects Agency (DARPA), conceived in February 1958 not only still exists, it has consistently made the US military the most advanced on Earth and unleashed life-changing technologies such as the internet, GPS and the computer mouse along the way.

Under the control of the Pentagon, DARPA has always maintained a low profile, but now journalist Michael Belfiore has written the first book about the agency. He spent time with the engineers charged with realising some very far-fetched ideas, which, if past performance is anything to go by, may also create society-changing spin-offs. He also gained unprecedented access to the agency's director, Tony Tether (now retired).

The current projects that Belfiore visits have typically ambitious goals - ones which dedicated New Scientist readers will be familiar with. He talks with two groups working to make prosthetic arms as nimble and light as the real thing, watches driverless cars work their way through real traffic in a bid to win a $2-million prize, and meets the creators of a portable robotic emergency room intended to keep injured soldiers alive long enough to reach hospital. He also learns about efforts to build scramjets able to race around the world in just a few hours.

Remarkably, DARPA doesn't own any of these labs. Its minimally bureaucratic three-level chain of command has gone essentially unchanged over the years. It works like this: the agency's director recruits a small staff of specialist programme managers to dream up far-fetched future technologies; the programme managers invite applications from engineers; the winning contractors frantically work to get results in the tight requisite time-frame of three to five years.

According to Tether, the most crucial step is the first. "The best DARPA programme managers, I swear, are science fiction writers."

Some of them are certainly fans of science fiction. A programme manager working on the robotic trauma-theatre tells Belfiore he got ideas for projects from sci-fi he had read in his 1950s childhood, and invited contemporary sci-fi writers to give seminars to help inspire DARPA engineers.

Striking the right balance between belief in the power of technology and credulity, though, is not easy. "You've got to fail some fraction of the time or you're not pushing the limits far enough," says Tether.

DARPA has certainly done its share of failing. Belfiore makes no mention that it was on Tether's watch in 2003 that DARPA started spending money on the junk-science notion of nuke-like "hafnium bombs" that would supposedly sidestep nuclear non-proliferation restrictions, although he does refer to the "considerable amounts" spent on psychic spying in the 1970s.

I would have liked more about how DARPA's ideas people handle the ever-present risk of stepping over the edge of what's possible. Unfortunately, the limited access granted to Belfiore wasn't enough to find that out, just as he wasn't told anything about the 50 per cent of DARPA projects that are classified. Belfiore does a good job of exploring the sunny side of the moon that is DARPA, but we still don't know what's lurking on the dark side.

Tom Simonite is the technology editor for NewScientist.com

If you would like to reuse any content from New Scientist, either in print or online, please contact the syndication department first for permission. New Scientist does not own rights to photos, but there are a variety of licensing options available for use of articles and graphics we own the copyright to.