Kelly Johnson was an aviation legend. The planes he designed ranged from the Lochkeed Hudson and P-38 Lightning to the Constellation, jet fighters, the U-2, and the SR-71. He was involved in dozens of designs and most were highly successful. He created and ran the legendary Skunk Works with an iron hand, literally known to kicking people in the ass on occassion. He was an honery SOB but a hell of an engineer. When Kelly Johnson retired, Ben Rich (of F-117A fame) took over the Skunk Works. He generally kept the Skunk Works intact and carried on its legacy. After Rich retired, another successor took over. The Skunk Works still exists and is still doing some good work like the F-35 Joint Strike Fighter.

Rutan's Scaled Composites operates following the Skunk Works model (Boeing has their Phantom Works organized along the same lines as well). The classic Skunk Works was established in 1943 to build the XP-80 jet fighter. Johnson wasn't allowed to interfere with existing production, so he grabbed about 30 of his most talented engineers and fabricators and isolated them from the rest of the plant. He didn't allow interruptions, unnecessary paperwork, and maintained rigid security. From the time Johnson signed the XP-80 contract to the first flight was an astonishing 143 days. They've repeated that success many times over the years, proving it wasn't a one time thing.

I've heard that fewer than 20 people were involved in building both the White Knight and SpaceShipOne at Scaled. Like Johnson, Rutan has hired some talented young engineers and fabricators*, given then his vision and leadership, and allowed them to do great things. Rutan is truly a visionary (he isn't even in the same building as the box, much less thinking inside one) who is helping nuture the next generation of people in his mold. If I were a young aeronautical engineer, I couldn't imagine a more stimulating and challenging environment to work. A few years there would be more valuable than any amount of graduate work, IMO.

Software and models can get you in the neighborhood of a good design, but it takes actual testing to make it work. The test results can be fed back into the models for validation purposes, but there's no substitute for testing actual hardware, both in wind tunnels and in flight. As Gary has pointed out, we've just about abandoned our ability to do many crucial forms of testing. In farming, they'd call that "eating your seed corn." Once eaten, you have nothing to plant for the next season, so you'll be out of business.

*One of the things Burt requires of all his engineers is that they can't design any part that they themselves can't build. This ensures the parts can be built and maintained. Unfortunately, union rules prevent that kind of thing at places like Lochkeed and Boeing. That shows, especially when it comes to maintenance.