Jeff,
	I am very upset by some of the things that happened yesterday.

	We have been working extremely hard at keeping news up and
running on the maxine that we had.  However, the hardware was inadequate
for the load.  This is evidenced by the number of errors we have been
getting on the disks, and by the fact that when news died yesterday, the
last recorded load was 255.27, with over 300 queued up connections.

	The number of people reading news has greatly increased.  We
have been providing a lot more news to many more people than we ever
were before.  This increase preceded Resnet, but is, and will continue
to be, aggravated by it.

	Left on its own, news service would have died sometime early in
the fall.  We have been keeping it going since then by putting many,
many hours into coming up with lots of optimizations and software hacks,
each of which bought us at most another month of time.

	At the time of this outage, we were down near the bottom of our
bag of kludges.  However, we were still working quite hard at coming up
a solution, and had been for several hours.  In fact, shortly before you
made the suggestion we ended up putting into production, a similar thing
had been suggested by someone else, and was among the possibilities
being discussed.  The rest of your suggestions were things we had
already tried, or that we had discarded as unworkable because we knew
more about what was going on.

	Even without new hardware, we had news up in a few hours.
Possibly, if we were sufficiently clever, we could have limped along
indefinitely, although it would have resulted in greater and greater
lossage for those trying to read news.

	Your stating that news died because the maintainers were
incompetent or lazy is not only incorrect, it is extremely offensive.
The way that you have not not talked about this at any time with any of
the active news maintainers makes it doubly so.  If you have a problem
with the way that news is being handled, or a suggestion for how things
could be improved, send mail to usenet@mit.edu.  News maintence takes a
lot of our time, is very stressful, and is totally thankless.  We don't
need IS unjustifiably deriding our competence on top of that.

	Moreover, I am extremely concerned by your unwillingness to
listen to us at all until things totally break.  When we first discussed
getting new hardware for the news server, we were told to write a formal
proposal and to submit it to you.  So, we did.  And we were told by you
that you didn't have time to deal with it then.  And whenever someone
tried to follow up personally on this, they were told you were either
out of town or too busy to deal with this.  However, no one else seemed
to have the authority to discuss our requests.

	News was shutdown yesterday because was just going to die when
it came up, and this was a cleaner way for the clients to handle it (ie,
they would give up and die rather than hanging).  It was not taken down
to make IS notice how bad things were and give us new hardware.  Having
someone respond when we went through the proper channels would have
avoided the outage yesterday for the users, and would have allowed us to
do a smooth cutover to to the new machine.