Newsgroups: comp.lang.smalltalk,comp.lang.lisp
From: cyber_surfer@wildcard.demon.co.uk (Cyber Surfer)
Path: cantaloupe.srv.cs.cmu.edu!das-news2.harvard.edu!news2.near.net!howland.reston.ans.net!pipex!demon!wildcard.demon.co.uk!cyber_surfer
Subject: Re: What makes Smalltalk different from Lisp?
References: <3ea2q6$kg8@freenet.vancouver.bc.ca> <ANDREW.95Jan4133201@srsunc.shlrc.mq.edu.au> <pdlogan.26.0009779F@orglobe.intel.com> <Pine.A32.3.91.950104163025.39742N-100000@swim5.eng.sematech.org> <patrick_d_logan.37.00096D8C@ccm.jf.intel.com>
Organization: The Wildcard Killer Butterfly Breeding Ground
Reply-To: cyber_surfer@wildcard.demon.co.uk
X-Newsreader: Demon Internet Simple News v1.27
Lines: 55
X-Posting-Host: wildcard.demon.co.uk
Date: Sat, 7 Jan 1995 14:34:11 +0000
Message-ID: <789489251snz@wildcard.demon.co.uk>
Sender: usenet@demon.co.uk
Xref: glinda.oz.cs.cmu.edu comp.lang.smalltalk:19317 comp.lang.lisp:16308

In article <patrick_d_logan.37.00096D8C@ccm.jf.intel.com>
           patrick_d_logan@ccm.jf.intel.com "Patrick D. Logan" writes:

> BTW what about SOAR (Smalltalk On A Risc)? How much of an OS was there? 
> Smalltalk is very nearly an OS anyway, right?

The Xerox PARC people used to say, "An operating system is a collection
of thing that don't belong in a language. There shouldn't be one."

Smalltalk 80 _was_ an OS. Not just an enviroment, as that could
be done, and these days is often done, on top of an OS.
 
> My long-term vision is that there will be an OS developed in Smalltalk or 
> similar language that will be very popular. As with the Lisp machines, there 
> will be C, Pascal, Ada, Fortran, etc. compilers for backward compatibility 
> with the outdated, brittle, C-machine systems and apps. (But as with the 
> Lisp machine, they compile to Smalltalk or whatever HLL.)

I'm not sure that such a pure approach would be popular enough these
days. It's still an attractive idea, and a good one for a long term
vision. I've just no idea when it might become popular again.

One thing that might do it, I suspect, could be MISC processors. That
could make a "custom" instruction set practical, and cheap, esp if we
don't demand a complete machine built from the chip upwards to be a
dedicated Lisp machine. I'd like to see something at the lower end,
like a plug-in card. There used to be cards like that, and I hope that
MISC could make that possible again someday.
 
> But who's going to fund such a thing? Maybe someone has to do it the way the 
> Linux community has done it. Maybe the approach should be to start with Linux 
> and a Smalltalk/Scheme/Dylan/? compiler and work backward, replacing Linux C 
> code with one of those.

I think that would have a greater chance of success (but what do I
know?) than expecting something like a Symbolics machine. Also, I
don't want a $120,000 machine! I'd like to see a machine that costs
the same as a typical desktop machine, or at least no more than a
low end workstation.

If that means going for a software-only route, then fine. That could
perhaps be done just as you've described it: start with something that
works and is available _now_, and then work toward your goal one step
at a time, with a working usable system at each step. It could also
be cheap enough not to become an ivory tower "luxury" item If it
like Linux where you'd only pay for the distribution, then it would
be _very_ cheap. That might be important. If people could try it
just out of curiousity, by borrowing a CD-ROM from a friend, then
more people would see that it was possible, and even practical.

I have a fondness for software...
-- 
CommUnity: http://www.demon.co.uk/community/index.html
CommUnity: ftp://ftp.demon.co.uk/pub/archives/community
Me: http://cyber.sfgate.com/examiner/people/surfer.html
