Read the rest
1. Visions not goals
2. Fund people not projects — the scientists find the problems not the funders. So, for many reasons, you have to have the best researchers.
3. Problem Finding — not just Problem Solving
4. Milestones not deadlines
5. It’s “baseball” not “golf” — batting .350 is very good in a high aspiration high risk area. Not getting a hit is not failure but the overhead for getting hits. (As in baseball, an “error” is failing to pull off something that is technically feasible.)
6. It’s about shaping “computer stuff” to human ends per the vision. Much of the time this required the researchers to design and build pretty much everything, including much of the hardware — including a variety of mainframes — and virtually all of the software needed (including OSs and programming languages, etc.). Many of the ARPA researchers were quite fluent in both HW and SW (though usually better at one than the other). This made for a pretty homogeneous computing culture and great synergy in most projects.
7. The above goes against the commonsense idea that “computer people should not try to make their own tools (because of the infinite Turing Tarpit that results)”.