News for hackers. Stuff that matters.
Re:Cosmonaut Crashed Into Earth 'Crying In Rage' [in 1967] by eliasmacpherson on Saturday March 19, 2011 @06:05AM (#2343280)
If you haven't seen it here's R. P. Feynman's Personal observations on the reliability of the Shuttle. I think the reason that additional safeguards were added was not because they begun to feel emotional about it, but rather NASA had underestimated the failure rate in the first instance. The emotional factor is what prompted the investigation.

http://science.ksc.nasa.gov/shuttle/missions/51-l/docs/roger...

Re:Cosmonaut Crashed Into Earth 'Crying In Rage' [in 1967] by gvb on Saturday March 19, 2011 @09:00AM (#2343474)
Appendix F is a must read. The tl;dr summary is:

Engineers at Rocketdyne, the manufacturer, estimate the total probability [of catastrophic failure] as 1/10,000. Engineers at Marshal estimate it as 1/300, while NASA management, to whom these engineers report, claims it is 1/100,000. An independent engineer consulting for NASA thought 1 or 2 per 100 a reasonable estimate.

— Personal observations on the reliability of the Shuttle by R.P. Feynman

Empirical post script (as of STS-133 24 February 2011)

  Total launches 	132
  Failures 		2
  Successes 		130
  Failure rate: 	1.5%
Edit: add opinions:

See also http://en.wikipedia.org/wiki/Rogers_Commission_Report#Role_o...

The reason for appalling and wildly inaccurate estimate of the failure probability is because the failure probability was a top-down "estimate." Someone Important apparently said "the space shuttle needs a reliability of 1/100,000". This mandate was pushed down on all the suppliers. The result was that the probability of failure was divided out onto all the systems, subsystems, subsubsystems... ad nausium all the way down to the bolts. The result was that all the subsystems had reliability numbers that could not be refuted (e.g. bolts at 1/10^12 IIRC), but could not be realistically substantiated either. Unfortunately, some of the probabilities of the major subsystems (SRB and ice damaging the heat shield tiles) were not accurate. The most accurate estimate was the engineer with a lot of experience in extremely complicated systems who gave an experience-based estimate, not a top-down (dictated) numerically derived estimate.

Re:Cosmonaut Crashed Into Earth 'Crying In Rage' [in 1967] by dmfdmf on Saturday March 19, 2011 @02:27PM (#2344303)
I think Feynman also identified how rationalizations due to schedule or political pressure were creeping in to safety assessments. When o-ring erosion was initially found they examined it and claimed that since the burn through was x%, that implied they still had y% of "margin" and used that to justify continuing flights. But this was never how the o-rings were designed to operate, 0% burn through is what was expected and now they were using engineering safety margin as operational margin, on a process (burn through) that was not fully understood.