response 4 of 7:

Jul 18 01:37 UTC 2008

[Interesting note: this is the 100th item in b/i/t/s....]

I'm not sure I really care about Linus's opinion, other than I mostly find
it annoying.  Mostly because I find him kind of annoying.  Mostly because I
find him about as arrogant as Theo Deraddt.  Which is pretty annoying.  The
guy wrote a kernel, that doesn't make him da Vinci.  But I digress....

The issue with security bugs is that they allow other people to influence
your system with malicious intent.  Thus, I disagree that they aren't more
important, because history has shown that there are people who can, and
will, exercise malicious intent and exploit those bugs to destroy or corrupt
your data.  That's bad.  Sure, it's bad when a buggy disk or application
crashes my machine, but there's nothing overtly malicious about it.

resp:3 I sort of disagree, as for a while, one of the BSDs (I think it
might have been OpenBSD) was releasing security advisories about ported
applications.  Maybe it was FreeBSD; I honestly cannot remember.  FreeBSD
certainly seems like the premier distribution in the BSD Unix world.

response 6 of 7:

Jan 7 06:31 UTC 2009

I inherited this freebsd mail server.  It seemed really slow so I got
and ran iozone.  It was able to produce something like 600 k bytes/ sec
sustained IO.  The same hardware in the adjacent rack slot running linux
and iozone was able to sustain about 55 M bytes/ sec.  So I migrated the
server functions and loaded linux to the freebsd hardware.  It was then
able to match the first linux system in performance.

To me a factor of 100 performance is a pretty important bug.  This
server was barely able to handle 5 simultaneous deliveries under qmail.
 Now it handles 50 simultaneous deliveries and acts like it is doing
zero.  Maybe freebsd should back off a notch on the security thing until
they get the basics working.

Other than things like that security is very important.  But then I'm
not a fair judge on that topic.  ;)