POS, OOFS, CL v Scheme, etc.

[Marc Wachowitz]

Brian Spilsbury <zhivago@iglou.com> wrote:
> I think we need a new, but very small dialect of lisp. One simpler
> yet more reflective than CL (and CL is reflective enough to mutate into CLOS
> for example). One capable of supporting a CL implementation over the top
> or a scheme implementation.
[...]
> Anyone willing to hack up a simple (and probably broken) blueprint
> for such a beastie as they see it?

Unless someone else wants to do it, or people tell me they'd not be
interested in me doing it at all, I'm willing to do it, and as long as
my writing style is sufficient, serve as editor for the evolution of
the result, in case this effort lasts (you'll definitely want some-
one else for an "official" report, though). I've pondered these things
for some time (about general language design for several years with
varying intensity, and more concentrated towards a Lispish core in
the recent months), and have read quite a few papers on that topic. As
I'm on vacation (at home) this and the next week, I'd invest most of
that time to get a very rough design outline within about a day, a more
concrete description, incorporating early feedback, perhaps until the
weekend, and a reasonably detailed document during the next week. Later
I'd be able to work at least an hour on it on most days, usually longer,
and on most weekends.

Generally, the idea would be to start somewhat more low-level than what
Henry Baker suggested in his remarks on DIN Kernel Lisp, but clearly
above register allocation and bit-level layout for most data (except a
few basic numeric types, but even that may belong into a kind of core
library), though such facilities should be available for implementation-
specific work.

This might also serve as a semantic layer for the VM people.

Comments?

In case someone missed it:
ftp://ftp.netcom.com/pub/hb/hbaker/CritLisp.html

-- Marc Wachowitz <mw@ipx2.rz.uni-mannheim.de>