Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP
Path: utzoo!utgpu!water!watnot!watmath!clyde!cbatt!ucbvax!hplabs!hplabsc!kempf
From: kempf@hplabsc.UUCP
Newsgroups: comp.lang.lisp
Subject: Re: LISP floating point performance
Message-ID: <1381@hplabsc.UUCP>
Date: Wed, 4-Mar-87 12:35:46 EST
Article-I.D.: hplabsc.1381
Posted: Wed Mar  4 12:35:46 1987
Date-Received: Fri, 6-Mar-87 21:44:53 EST
References: <16921@glacier.STANFORD.EDU> <441@spar.SPAR.SLB.COM> <16930@glacier.STANFORD.EDU>
Organization: Hewlett-Packard Laboratories
Lines: 27
Summary: Not "don't", just "might consider the alternatives"

In article <16930@glacier.STANFORD.EDU>, jbn@glacier.STANFORD.EDU (John B. Nagle) writes:
> 
> floating-point computation.  The notion that LISP users don't do floating point
> work is decidedly out of date.
> 
> 
> 					John Nagle

The intention of my posting was not to imply that Lisp users "don't"
do floating point, just that, given modern foreign function capabilities
and the existence of tried and true reusable software like the IMSL
libraries, it may be more cost effective to do the floating point
using them rather than trying to recode standard algorithms in
Lisp. The inclusion of hooks for good floating point into Common Lisp
certainly leaves open the possibility for incorporating floating
point optimizations into Common Lisp compilers.

I noticed that the listed applications using heavy Lisp floating
point all seemed to be prototypes (please correct me if I'm wrong).
Are there any Common Lisp applications that are products which
use heavy Lisp floating point? If I were in the position of
developing a product requiring both symbolic computation and
floating point, my tendency would be to go with a tested
reusable module like the IMSL library for standard floating
point computations, and use Lisp for the symbolic part.

		Jim Kempf	kempf@hplabs.hp.com