> at some point, one needs to launch the thing, despite the complains There's a big difference between "complaints" because something is not optimal, and warni

> at some point, one needs to launch the thing, despite the complains There's a big difference between "complaints" because something is not optimal, and warnings that something is a critical risk. The Thiokol engineers' warnings about the O-rings were in the latter category. And NASA knew that. The summer before the Challenger blew up, NASA had reclassified the O-rings as a Criticality 1 flight risk, where they had previously been Criticality 1R. The "1" meant that if the thing happens the shuttle would be lost--as it was. The "R" meant that there was a redundant component that would do the job if the first one failed--in this case there were two O-rings, primary and secondary. But in (IIRC) June 1985, NASA was told by Thiokol that the primary O-ring was not sealing so there was effectively no redundancy, and NASA acknowledged that by reclassifying the risk. But by the rules NASA itself had imposed, a Criticality 1 (rather than 1R) flight risk was supposed to mean the Shuttle was grounded until the issue was fixed. To avoid that, NASA waived the risk right after reclassifying it. > at some point, one needs to say "yes" and take risks, otherwise nothing would be done Taking calculated risks when the potential payoff justifies it is one thing. But taking foolish risks, when even your own decision making framework says you're not supposed to, is quite another. NASA's decision to launch the Challenger was the latter.