6 Sep
2022
6 Sep
'22
11:13 p.m.
On Tue, Sep 06, 2022 at 11:07:19AM -0400, Douglas McIlroy wrote:
Maybe this one?
B.P. Miller, L. Fredriksen, and B. So, "An Empirical Study of the Reliability
of UNIX Utilities", Communications of the ACM 33, 12 (December 1990).
http://www.paradyn.org/papers/fuzz.pdf
(Research)
Unix ... 'shipped' with zero known bugs.
It wasn't a Utopia. Right from the start man pages reported BUGS,
though many were infelicities, not implementation errors.
Dennis once ran a demo of a ubiquitous bug: buffer overflow. He fed a
2000-character line on stdin to every program in /bin. Many crashed.
Nobody was surprised; and nobody was moved to fix the offenders. The
misdesign principle that "no real-life input looks like that" fell
into disrepute, but the bad stuff lived on. Some years down the road a
paper appeared (in CACM?) that repeated Dennis's exercise.