I just looked at Henry Baker's "Critique of DIN Kernel Lisp Definition
Version 1.2": what has happened in the meantime?
It'd be interesting to have a lisp that is good for numerical
simulation running on current (and future) massively parallel
machines. How about Thinking Machines's lisp? Anything for a
Cray T3D? (Or E?) Or must `Grand Challenge' projects suffer through
Fortran and C++?
Regards.
--
Mike Rilee NASA/GSFC Mailstop 930.0 Greenbelt, MD 20771
······@hannibal.gsfc.nasa.gov Ph. (301)286-4743 Fx. (301)286-1777
--
Composed using Oberon. http://www-cs.inf.ethz.ch/Oberon.html
Mike Rilee <······@hannibal.gsfc.nasa.gov> wrote:
> I just looked at Henry Baker's "Critique of DIN Kernel Lisp Definition
> Version 1.2": what has happened in the meantime?
DIN Kernel Lisp has much evolved since that time: An object system has
been added. It has been merged with Japan Kernel Lisp; however, I don't
know whether the spirit of Japan Kernel Lisp has been preserved in this
process. Since then, it has been proposed as ISO ISLisp (see
ftp://ftp.harlequin.co.uk/pub/kmp/iso/v19/). It's going to become an
ISO ISLisp standard sometime soon now.
> It'd be interesting to have a lisp that is good for numerical
> simulation running on current (and future) massively parallel
> machines. How about Thinking Machines's lisp? Anything for a
> Cray T3D? (Or E?) Or must `Grand Challenge' projects suffer through
> Fortran and C++?
Unfortunately, there aren't many ISLisp implementations out there.
For numerical simulations, I'd recommend CMU Common Lisp. It produces
machine code which is equal to in quality, sometimes superior, to
C/Fortran code. (See also the excellent article by Richard Fateman et al.
in Math. Comp. or SIAM J. Comput. (?) two years ago.) CMU Common Lisp
runs on Sparc, Mips, [3456]86, HP, DEC Alpha systems, so you would better
stay away from non-standard architectures.
Note that one of your collegues, Erann Gat, is already using Common Lisp
for a space-flight project.
Thinking Machines having gone out of business, I don't think *lisp is
still supported.
Best regards,
Bruno
Bruno Haible email: <······@ilog.fr>
Software Engineer phone: +33-1-49083585