What DARPA's Been Up To, At Length 54
The New York Times takes an inside look at DARPA, the secretive defense agency, mentioned frequently on Slashdot, that is "changing the way we use machines — and the way they use us" in the form of a review of Michael Belfiore's The Department of Mad Scientists. Besides tracing the history of the agency, Belfiore's book expounds on the well-known Grand Challenge and its link to ever-more-automated vehicle control in civilian and military contexts, as well as other DARPA pet projects, including robotic surgery, information analysis, and the integration of electronics with the human body.
Re:The suckitude that was DARPA head Tony Tether (Score:2, Informative)
With an incurious, aggressive president, we got an incurious, aggressive DARPA head, who cut long-term and academic research in favor of short-term corporate research.
I agree as long as we make the translation, "long-term and academic research" == useless research and "short-term research" == useful research. It's worth noting that Anthony Tether headed DARPA over the period when DARPA became popular on Slashdot due to its much cooler projects. To be blunt, for all the talk of the value of academic research, it really isn't that useful or interesting. I'm sure academics are overjoyed to be able to hog the public fund trough again, but that doesn't mean that they deserve the money or that it's a better use than the impressive projects of the Bush era.
Reading on, you cite the ARPAnet as an example of a long-term project. The problem is that the project was short-term. Government needed a network that could survive a nuclear blast. They got that within a few years. They didn't have any interest in the rest of the stuff that went on the network. It was a short-term project with long-term benefits that just happened to use academia as the means rather than businesses. We can't use it to claim that funding long-term/academic research is something that DARPA should be doing.
Tether did a good job (Score:4, Informative)
Tony Tether (whom I've met) did a reasonably good job with DARPA. Especially in robotics. He was behind the DARPA Grand Challenge, which was done partly to give academic robotics departments a serious butt-kick. Academic robotics had been funded by DARPA for decades, but nothing fieldable was coming out. The reason that major universities devoted entire departments to the Grand Challenge was that DARPA had told them quietly that if they didn't do well, their funding was going away. Prior to the Grand Challenge, a typical academic robotics project was one professor and a few grad students producing a thesis on an obscure topic. Universities weren't organized to do system integration and make all the subsystems play together. Now they are.
It was time to cut back on Government-funded R&D in computer science, because it's a mature technology. DARPA shouldn't be funding "high performance graphics" - industry, Hollywood, and the game industry are doing that just fine. Networking is in good shape. DARPA hasn't been influential in operating systems since the 1980s. DARPA never had much of a role in personal computing at all.
DARPA isn't the NSF. Their job is to develop technology DoD can use.