On Sun, 2 May 1999, Curt Sampson wrote:

# Anyway, Charles was dead right on this one; Greg can set up his
# CVSup server and if people use it, great; we can have a flamewar
# about putting a CNAME for it under netbsd.org. If nobody uses it
# we just don't get to have that flamewar. I'm sure we can come up
# with something else to fight about anyway. :-)

While I see that it was intended in humour, I think I might put
a "* 0.5" after that, because you've made a point here, intentional
or not.

I'll admit to being one of the increasingly vocal crowd on some issues,
and I acknowledge that some of my rants have rendered my voice invalid
in some eyes (which is a shame, but that was at least half my own doing).
That notwithstanding, I'd hate to think that it boils down to the concept
that the only thing that these lists are good for is electronic char-fests
with appropriate sacrifices to the fire gods Kingsford and Weber.

In truth, I see a lot of good information go back and forth, and believe
me, I know how difficult it is to keep one's emotions in check regarding
certain issues (rc.* vs rc*.d/<ordered-file-list>, anything remotely
System V-ish vs. The One True (Berkeley) Path, namespace redefinition).
This much is blindingly obvious, and we have some heated discussions.

It's when we descend into name-calling or electronic pixoventriloquism
[typing words into other people's mouths] or when the debate seems to go
nowhere after a while, yet we still flog the deceased equine as though
we expect it to rise, phoenix-like, from its own rubble, that it becomes
obvious that the energy would be better spent elsewhere.

It helps to humble oneself before the fact that opinions are like
recta:  Everyone has one but nobody wants to look at the other guy's.

Regarding runlevels/machine states:  If we can do it in a way which
is maintainable yet more reasonable than the rc*.d/[SK][0-9][0-9].+
manner of SysV, this would be a reasonable thing to do.  Otherwise,
we should leave it alone.

Regarding namespace definitions, namely wrt cron:  We will probably do
what makes the most sense for our environment, on an individual basis.

Regarding shells:  I still vote for csh by default for root, but I think
that's just my own dinosauric tendencies showing through.

Regarding CVSup:  I had my own arguments against it, but if it's available
for the SPARC platform, I'll probably use it because pulling deltas over
at 28.8 would be a much nicer way than pulling over files wholesale.

Regarding updates in general:  I seem to remember sometime around the
last release that times on all files were updated in such a manner so
as to create a problem in which a 'sup -o' would update the entire
source tree even though little to nothing had changed.  Can we expect
this to happen again?

I will remind that what I have posted is purely opinion, and I *do*
try to keep my vitriol in a locked cabinet :-).  I will acknowledge,
in this same breath, that I am a dinosaur in thought sometimes --
change comes slowly and sometimes painfully for me.  That aside, I suspect
that I am not the only one for whom this is true.

Apologies for one's opinion or fundamental expression should never be
needed.  How one presents that opinion or expression may be a different
manner, however, and in that respect, though I may be dismissed as an
ignorant and foolish individual, I wish to convey apologies for any
friction I have caused by way of caustic phraseology, although I will
note for the record that the higher up on the intellectual ladder one
is, the more socially inept one is likely to be (statistically speaking).

We now return to our regularly scheduled discussion, already in progress.

				--*greywolf;
--
NetBSD: Network Your World.