[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
- To: JLK at MIT-MC
- From: Howard I. Cannon <HIC at MIT-MC>
- Date: Sun ,13 Jan 80 19:20:00 EDT
cc: BUG-LISPM at MIT-MC, Agre at MIT-AI
Thanks for the public flailing -- being a masochist I truly appreciate it!!
The system was designed around the presumption that the Chaosnet does
not go down arbitrarily, and that a process ca get "interrupts" when
a Chaosnet packet arrives. These two premises are true on all other systems
that support the Chaosnet. Unfortunatly, the Lisp Machine is not up to the
task of doing this -- no one turns on sequence breaks because they are
afraid everything will break (possibly a justifiable concern), and I can't
be expected to make the system be perfect under all possible circumstances.
(Next you will be asking me to make sure it dies gracefully when someone
pulls an arbitrary chip out of the processor -- it's VERY MUCH THE SAME
THING. The system is expected to operate well under constraints which
are different from those that it was designed for.) I will fix bugs of
this nature that arise when sequence breaks are on -- otherwise, I will
not WASTE my time making the system robust under unrealistic conditions.
The Lisp Machine as it currently stands IS A MARGINAL COMPUTING ENVIRONMENT,
and all users of it better understand that. It's probably the best
environment there is in certain respects, but in others...we are
doing our damndest to fix it, but arbitrary flames to BUG-LISP, JLK, about
problems that I have already explained to you, is irresponsible and
wastes all of our time.