Article-I.D.: aero.4201
Posted: Mon Nov  3 16:41:40 1986
Date-Received: Wed, 5-Nov-86 06:30:50 EST
References: <12464@vuwcomp.UUCP> <2501@osu-eddie.UUCP>
Reply-To: coffee@aero.UUCP (Peter C. Coffee)
Organization: The Aerospace Corporation, El Segundo, CA
Lines: 30
Xref: watmath comp.sys.mac:1 misc.wanted:140

In article <2501@osu-eddie.UUCP> verber@osu-eddie.UUCP (Mark Verber) writes:
 > There are four 'real' Lisp compilers for the Macintosh...
:
 > #2. ExperLisp.  I don't recommend this compiler.  The compiler is
 >     buggy.  They advertise that it is CommonLisp compatible but
 >     it doesn't support closures, use lexical scoping or have
 >     multiple name spaces.  I guess you can call it CommonLisp
 >     since the function names are the same.  I have talked to a
 >     number of Lisp implimentors about the product, and all have
 >     felt that the ExperLisp compiler just wasn't that hot.

I regularly work on a Symbolics 3640 and on MS-DOS machines as well
as on a MacPlus with ExperLisp. ExperLisp 1.5 is quite stable, with
generally higher speed than earlier releases; it's usually about triple
the speed of interpreted GCLisp on a 7MHz 8086. The user interface is
sublime. It does _not_ produce stand-alone applications, but if you
can wait until early December they may have a surprise for you (a non-
disclosure agreement prevents me from saying more). BTW, ExperLisp has
_always_ been lexically scoped, even though this put severe constraints on
lambda expressions in earlier releases: 1.5 fixes these. On a 1 MB Mac
Plus, I find that ExperLisp 1.5 with 24000 cons cells and 4800 symbols
allocated runs from 1-3% the speed of compiled Common Lisp on a 2 MB
Symbolics 3640.

I have no commercial relationship with ExperTelligence, though they have
expressed an interest in marketing my on-line documentation system
for ExperLisp (still in development) as an enhancement product. If I
didn't like ExperLisp, however, I wouldn't bother...

Regards, PC