Last week it was bandied about this group about possibly creating a
lisp-os/lisp machine. Now, I'm *very* new to lisp and couldn't do much on that
side of the beast, but I do know a bit about operating systems. As a result,
I'd be willing to do some work towards making a Mach or NetBSD based home for
such a lisp on Intel machines. But there are a few questions we should think
about first.
In essence, such a system would be Unix at it's core. Lisp would become the
command shell (lsh(1)?) that would be the direct interface to the system. Ah,
but what type of lisp - one that byte compiles ala Clisp, or one that compiles
to C, ala AKCL/GCL? Also, which code base is the cleanest (for addition of
threading and other such desireable goodies)? Additionally, how far from CL is
Elisp? Emacs itself could be the "shell", as it were, if the differences are
too significant. Finally, and to me the most important question, is how close
do we want to be to the standards (such as they are)? Do we want to build a
"Production" system or a "Research" system? This could be a trivial point, but
I think that if something like this is going to be done, it needs to be
coherent in it's philosophy from the beginning.
These are just some initial thoughts. If anyone is interested in this, let's
see what we can put together.
William Lewis
····@macc.wisc.edu
(*standard disclaimers*)
In article <··········@news.doit.wisc.edu>, ····@MACC.WISC.EDU (STEB) writes:
> Last week it was bandied about this group about possibly creating a
> lisp-os/lisp machine. Now, I'm *very* new to lisp and couldn't do much on that
> but what type of lisp - one that byte compiles ala Clisp, or one that compiles
If something like this was to be done, I would go for it and create a really
clean object-oriented micro kernel OS with a MOP-based Scheme as the native
language.
> do we want to be to the standards (such as they are)? Do we want to build a
> "Production" system or a "Research" system?
I would say a "Reseach" system with a definite "Production" vocation
(after all, Unix was a research system, and C was invented concurrently;
what's more, MOP-based scheme compilers are still very much at this
stage).
--
F.D. Mato Mira
Computer Graphics Lab ········@epfl.ch
EPFL FAX: +41 (21) 693-5328
In article <···················@aye.cs.nyu.edu>, ·····@graphics.cs.nyu.edu (Matthew Fuchs) writes:
|> In article <··········@disuns2.epfl.ch> ········@di.epfl.ch (Fernando Mato Mira) writes:
|>
|> In article <··········@news.doit.wisc.edu>, ····@MACC.WISC.EDU (STEB) writes:
|> > Last week it was bandied about this group about possibly creating a
|> > lisp-os/lisp machine. Now, I'm *very* new to lisp and couldn't do much on that
|>
|> > but what type of lisp - one that byte compiles ala Clisp, or one that compiles
|>
|> If something like this was to be done, I would go for it and create a really
|> clean object-oriented micro kernel OS with a MOP-based Scheme as the native
|> ......
|>
|> A couple of years ago, Suresh Jagannathan of NEC Research was working
|> on a project called STING which used Scheme as both the O/S and
|> development language for a distributed computing system. I don't know
|> the current state of that project.
|>
|> Matthew Fuchs
|> ·····@cs.nyu.edu
There are a number of papers describing Sting that may be retrieved via
anonymous ftp from ftp.nj.nec.com in the directory /pub/pls.
A WWW URL should also soon be available.
-- Suresh
In-reply-to: ······@research.nj.nec.com's message of 17 Jun 1994 12:59:14 -0400
>>>>> "Suresh" == Suresh Jagannathan <······@research.nj.nec.com> writes:
In article <··········@swifty.nj.nec.com> ······@research.nj.nec.com (Suresh Jagannathan) writes:
Suresh> There are a number of papers describing Sting that may be
Suresh> retrieved via anonymous ftp from ftp.nj.nec.com in the
Suresh> directory /pub/pls. A WWW URL should also soon be available.
In "A Overview of the STING Operating System" James Philbin says:
"The next step is to integrate Sting into the microkernel such as Mach
or Chorus"
More recently, in "High-Level Abstractions for Efficient Concurrent Systems"
you write:
"The physical machine configuration maps physical processors to
lightweight Unix threads; each note in the machine runs one such thread"
Has Sting been integrated with a microkernel?