In support of (and in sympathy with) Berkeley

[cak at PURDUE.ARPA]

From:  Christopher A Kent <cak at PURDUE.ARPA>

Since my mini-flame seems to have precipitated this whole series of
Anti-Berkeley flames, I want to make a few things clear.

First off, I write these things with the usual personal opinion
caveats; they represent my view and my view alone. Period.

Secondly, I have an incredible amount of respect and praise for Bill
Joy, Sam Leffler, Kirk McKusick, Mike Karels, and the cast of
(apparently) thousands that have made 4.2 possible. Without the code
that they have produced over the life of the DARPA project, I would
probably still be running 32/V, or worse, be living on a pdp-11. V6 was
nice, but.

My comments came after a particularly arduous session of debugging a
particular portion of my 4.2 release, only to look at my mail and find
that someone had just re-discovered some fixes I had done a week ago,
and that they had fought my current battle a week ago. 

I probably fell into the trap of thinking of Berkeley as a support
organization for just about long enough to type that flame and send it
off. [Mea Culpa]*1000. It was wrong, and I admit it. I sit in a
situation similar to theirs and the one described by Randy Frank,
except on a smaller scale, and I understand. I was off base.

However, much of what I say still holds true. I don't have a copy of my
note, but perhaps with minor changes, much of it would hold through. I
don't want or need hand-holding; what I want is a good way for 4.2
licensees to be able to exchange bug reports and fixes in a timely
manner. It isn't fair to expect Berkeley to provide this; they aren't
being paid for it and don't have the manpower. I realize all that. But
there is still a need. The fixes seem to filter to Berkeley before
anywhere else; they do collect them and peruse them; it would be nice
if we could see the results of that action. This isn't currently
possible for the great majority of 4.2 licensees, and I'd like to see
that change. It is possible to get help out of UCB, but it is often
slow to come (understandably).

This is a whish list item, mind you; I'm not saying that it is what
SHOULD be, it is what would be NICE. If Berkeley can't devote the
manpower to it, perhaps they would be willing to let someone else do
so. A lot of people are out there testing combinations of hardware and
software that Berkeley can't possibly be expected to have on site, and
fixing problems that crop up; for licensing reasons, Unix-Wizards is
not an appropriate place to discuss many of these; we need another way.
The Usenix newsletter used to be a good forum for this; that seems to
have fallen by the wayside, too. (More about that in another flame
someday, perhaps).

I hope I have smoothed any ruffled feathers, cleared up my good name,
and made clear my original intentions. I am eternally glad I have 4.2;
AT&T doesn't seem to understand what a large number of Universities and
other DARPA contracters want, and Berkeley seems to. I'm glad we have a
choice, or I'd be cursing and pulling my hair out and writing TCP/IP
for SysV, just like a lot of other folks.

Cheers,
chris
----------