> "Challenger and Columbia were both spectacular failures, and both were the result of dysfunctional corporate culture at NASA."

Feynman's Appendix F should be required reading for every sort of engineer: http://science.ksc.nasa.gov/shuttle/missions/51-l/docs/roger...

It's a little bit long, and gets a bit technical at some points, but everyone should read it anyway.

Note however, that this example doesn't fit in at all with the contrarian view in this thread. Feynman's report was about accurately estimating the failure rate of the shuttle and about the bureaucratic culture that popped up to protect itself and ended up making that failure rate worse by not letting engineers do their jobs.

But if during the Challenger investigation an engineer or a manager had told Feynman "There are place and time for failure. Failure on a launch operation isn't one of them", he would have regarded them as completely incompetent and part of the problem.

The very nature of the enterprise is risky, and so the goal was to manage the risk of failure to an acceptable limit. Cover-your-ass culture can be improved, engineers can be allowed to do their jobs, but you aren't going to eliminate risk. The fact that we can even try to put a number like "1 in 50" or "1 in 100" on estimated space flight failure rates is exactly what blhack is talking about. We could judge that too high a rate, but if we did, we wouldn't be going to space at all, at least not anytime soon. That's the tradeoff.

Note that the engineers and especially the astronauts were well aware of that risk, and you can find almost exactly that sentiment echoed by them in interviews after the Apollo 1, Challenger, and Columbia accidents. Many astronauts were most worried that the American public couldn't stomach the idea that failure came with any ambition in spaceflight and so would shut the program down.