Re: compiler defects and mission critical failures

The ability of bugs to hide in the face of rigorous testing and
use is indeed almost unbelievable. When I worked for DEC on
FORTRAN, we found that every once in awhile, a user would report
a bug that had been there since day 1 (where day 1 was nearly 10
years earlier), and had eluded all the test suites, plus a
user community that numbered in the thousands. Oddly enough,
no matter how obscure the bug, it seemed that once it was
discovered by one customer, another in a different part of
the world would report the same bug within a week or two of
the first report. After all these years working on compilers,
and seeing the number of bugs in them, what's really amazing
is that ANYTHING works at all, given that most software has
to go through these beasts...


And now we return you to your usual technical discussions.
Mike
[This message, which was sent nearly six months ago, just showed up
here two days ago. Speaking of software bugs ... -John]
--

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