Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP
Path: utzoo!mnetor!seismo!rutgers!sri-spam!ames!ucbcad!ucbvax!cs.umass.edu!FORSTER
From: FORSTER@cs.umass.edu (David Forster)
Newsgroups: comp.sys.atari.st
Subject: LISP for the ST
Message-ID: <8703060629.AA10665@ucbvax.Berkeley.EDU>
Date: Thu, 5-Mar-87 09:41:00 EST
Article-I.D.: ucbvax.8703060629.AA10665
Posted: Thu Mar  5 09:41:00 1987
Date-Received: Sun, 8-Mar-87 03:16:40 EST
Sender: daemon@ucbvax.BERKELEY.EDU
Distribution: world
Organization: The ARPA Internet
Lines: 50

I'm looking for a decent version of LISP for the ST, but there doesn't seem to
be much out there.  I've only heard of Cambridge LISP, XLISP, and something
from Robinson Systems in the U.K.
  The version of XLISP which I have is too slow and buggy for my purposes,
but it's about a year out of date and I expect that XLISP has improved in the
mean time.  I'd fix the bugs I run into in this version, but I don't have a
C compiler, and have no intention of buying one just for this.  In any event,
there's still the problem of speed.
  The Robinson Systems LISP is simply inadequate, according to the material
they sent me.  It appears to have smaller coverage than does XLISP, and they
still want something like $150 for it.  It's supposed to be based on Common
LISP, but lacks such things as setf, defstruct, and even defmacro.
  All I have seen of Cambridge LISP is advertising by Metacomco, and a review
in AI Expert Magazine.  According to the review, there are some problems with
the product: it uses dynamic scoping when interpreted, but lexical scoping
when compiled; it seems to be slow*; it's not Common LISP, but Portable
Standard LISP; it doesn't have defstruct's, but it does have arrays**; as I
recall, it's a nuisance to compile functions, but I don't recall why.
On the plus side, it does have a compiler, it has VDI functions and such
defined, and it's possible to save images, meaning that binary code can be
distributed without problems.
* - Benchmarks showed that it ran 20 times slower than equivalent products on
    the Mac, BUT the Cambridge LISP was being interpreted, and the Mac stuff
    was running compiled.  If the Cambridge code had been compiled, it would
    probably have performed better, but 20 times better?  I don't think so.
**- Why on earth they didn't go that extra step to define structures I do not
    know.  Surely it's not that much work?  It should be possible to define
    structs based on arrays without very much work (and/or based on lists, but
    access to arrays will be faster).

Does anyone out there have information about other LISP's?  Does anyone have
a copy of Cambridge LISP which they'd like to report on?  I must admit that
I'll be finicky about the LISP I buy, since I'm used to working on LISP
machines like the TI Explorer and Symbolics 3600-series machines.  I don't
expect an ST LISP to match that level of performance, but I do expect it to
have most of the power of Common LISP and fairly similar syntax, and not make
me wait too much.  It ought to have a package system, structures, and that sort
of stuff, but there are work-arounds for those.  It needn't be cheap, but I
want good value for my money.

Sorry if this has been discussed before, but our site missed two months of
digests for some reason.  (If it has been discussed before, can someone please
mail me what they have on it?)

Thanks, David Forster

CSnet:   forster@umass-cs.csnet
ARPAnet: forster@umass-cs.csnet%csnet-relay.arpa
BITnet:  forster@umass
USENET:  your guess is as good as mine :-)