Re: Paul Graham's PyCon Keynote

[Guy Steele - Sun Microsystems Labs <address@hidden>]

   To: Eli Collins <edc206@cs.nyu.edu>
   Cc: ll1-discuss@ai.mit.edu
   Subject: Re: Paul Graham's PyCon Keynote
   From: Luke Gorrie <luke@bluetail.com>
   X-Sincerity: 14% (approx.)
   Date: 11 Apr 2003 03:19:43 +0200
   User-Agent: Gnus/5.09 (Gnus v5.9.0) Emacs/21.2
   
   Eli Collins <edc206@cs.nyu.edu> writes:
   
   > Joe Armstrong argued at LL2 that concurrent programs are easier to write
   > and understand than sequential programs. Could an Erlang descendant
   > running on commodity PCs with thousands of FPGAs finally liberate us from
   > the von Neuman Style? I'm not so sure parallelism can't climb to the top
   > of the ladder of abstraction.
   
   Have you looked at Connection Machine Lisp? I took the liberty of
   putting the paper on the internet,
   http://www.bluetail.com/~luke/misc/papers/ConnectionMachineLisp.pdf
   
   I was blown away by the novelty, and I'm very eager to hear some CM
   anecdotes from the Thinking Machines veterans on this list. I
   understand that CMs were applied to some interesting problems, but my
   searches haven't revealed any "case study" type of literature to say
   how it all turned out!

Thanks for the praise.  Let me comment that CM-Lisp turned out
to be harder to implement on a Connection Machine than we thought.
I would say that the main claim to fame of CM-Lisp might be
that it insired Guy Blelloch to design another language called
NESL, which was thoroughly implemented and users were able to
gain some real experience with it.

--Guy