Re: failure due to compiler?

bobduff@world.std.com (Robert A Duff) writes:


|> Yes, compilers have bugs. I've seen many. What would be more
|> interesting (to me, anyway) would be cases where compiler bugs actually
|> caused some real damage. E.g. so-and-so plane crashed because of a
|> compiler bug, and killed so-and-so number of people. Or so-and-so
|> company lost $100,000 because of a compiler bug. Or some such thing.


I imagine that a company losing $100.000 is actually fairly frequent.
I'm sure, for example, that we've lost more than that at my present
customer site due to compiler bugs. It usually takes at least a week of
debugging to determine that it is a compiler bug, and another week to
find a work-around. Then someone else falls into the same case, except
that it looks different superficially. In almost no time, on a large
project, you'll find that you've spent a man-year. And a man-year comes
out to about $100.000 (or a little more).


Even when you know the bug is there, it costs. Every time you come up
with a new design, someone has to spend time verifying that it isn't
going to run up against the bug. Sometimes, you end up adapted
definitly inferior designs just because you know they don't exercise the
buggy part of the compiler.
--
James Kanze Tel.: (+33) 88 14 49 00 email: kanze@gabi-soft.fr
GABI Software, Sarl., 8 rue des Francs-Bourgeois, F-67000 Strasbourg, France
--

Return to the comp.compilers page.
Search the comp.compilers archives again.