Re: How fast is that machine???

[Drew McDermott <Mcdermott@YALE.ARPA>]

  There is more to a computer than how fast it runs, just as the top
  speed of a car is of only minor importance compared to comfort,
  maneuverability, etc.  Before one rates a machine as "an almost
  unqualified WONDERFUL", one needs to ask some hard questions:

In my opinion, the 3600 (or any other Lisp machine) beats the Apollo
in all the other categories already, plus some you didn't mention.
For instance, the Apollo window package was apparently designed by
chimpanzees.

   1.  What is the cost per node for all that speed?  If a node that
   costs twice as much runs 3-5 times as fast, is that really a good
   deal, given that most of the time it sits practically idle doing
   text editing?

With an Apollo, the user sits there idle most of the time, waiting
for garbage to be collected.  However, it is probably true that a 3600
is inappropriate for text processing, and we would need some other
hardware for that.

The real cost-per-node issue is whether or not we would need a sign-up
sheet system to cope with the demand for scarce, expensive Lisp machines.

   2.  What is the true cost of switching systems, once you factor in
   the additional staff needed to support it and the increased difficulty
   of interaction with others in the local community?

This is a real problem, but it's inevitable, unless we can all stick
to Apollos forever.  I doubt that interaction with the community will be
a problem, since something like Common Lisp will be available on all
our machines in the not-too-distant future.  Common Lisp will be able to
do anything that T can do, just a lot uglier.  Also, remember that
Common Lisp will ease communication with people *beyond* the local
community.

   3.  How good is the software?  Does it support transparent remote file
   access?  Mail?  TeX?  Scribe?  Graphics?  Yale-style editors?  C?

Yes to all but word processors and C.  As I said above, word processing
would be done on Apollos or Apples.  No one would need C on a Lisp machine.
(Why would we want it?  Portability of system software?)

   4.  How good is the Lisp?  Does it have the clean semantics to
   facilitate the development of large systems that T has?  How long
   does it take to learn to use effectively (i.e. what is its "kludginess
   coefficient")?  How many of our faculty and students have the time
   to invest in learning to use it effectively?

The Lisp is Zetalisp, which will no doubt evolve into Common Lisp.  (CL
was heavily influenced by the design of ZL.)  As I said before, there is
no doubt that Common Lisp is "dirtier" than T.  But I don't think it
significantly harder to use than T.  The programming environment is
much better.

By the way, as ZetaLisp evolves, closures will get a lot more efficient.
If you use Nisp "closures," you probably pay no overhead now.

   5.  What is its growth path, i.e. if you take similar benchmarks in
   two years' time, what will you expect to see?  Is there any reason to
   believe the Symbolics class of machines will remain faster than the
   Apollo class?  Maybe I hold out too much hope for the new T compiler,
   but I had the impression that it might yield a speed increase of
   two to three times.  If so, the Tern running T would already approach
   the 3670.

I would expect Lisp machines to continue getting faster, as they have
in the past.

I am curious to know whether Terns are catching up with Lisp machines,
but apparently Apollo has no interest in our curiosity, since they have
made no effort to deliver a real Tern to us.

                          -- Drew

-------