This thread has brought out once again how the truly remarkable level of
safety we enjoy in our civil aviation system has evolved primarily
through the learning experiences of a long series of crashes and
accidents.

These accidents were individually tragic -- but also individually small
in some reasonable sense of that term, and thus acceptable.

It also seems to me they were in most cases largely unanticipated and
perhaps largely "unanticipatable" -- we had to have them, in order to
evolve to the level of safety we have today.

It's these aspects of aviation safety that bother me about the analogous
case of nuclear safety (in the sense of both nuclear power, and nuclear
weapons risks). We've had a few nuclear accidents, and undoubtedly
learned from them.

But we've not had the sustained chain of nuclear accidents to teach us
the risks and the necessary safeguards of nuclear technology -- and we
may never have them until it's way, way too late.

A worst case aviation accident (a fully fueled 380 falling out of the
sky onto a fully filled football stadium) might kill a few tens of
thousands. A worst case nuclear accident might kill or poison many
hundreds of thousands and upwards, and render a major metropolitan area
or half a state uninhabitable for decades to centuries.

And, as my wife keeps saying, "fail safe systems by definition fail by
failing to fail safe".