Tue, 29 Apr 1997 17:43:31 -0700
Fare Rideau wrote:
> Luca Pisati rightly pointed out:
> >> I'm just trying to understand what are we trying to do here.
> >> [follow 3 alternatives]
> Richard Coleman answered:
> > [...] I believe most people are interested in variations
> > of either (1) or (2). Options (3) is too ambitious and would probably
> > never be completed.
> Well, I happen to be interested in (3) [Radical new approach of an OS],
> but in the frame of another project, Tunes.
> I don't agree why it would _never_ be completed;
> well, I admit no one (including me)-; did complete it _yet_.
I admit I'm much more interested by option (3).
I was never interested by Unix, which is _old_ technology,
and even less by Windows (which is _old_ by definition, to me).
Since I had the opportunity to work for few years (also in Paris)
on Symbolics Lisp Machines, I think I had an idea of what a _new_
OS could be. For sure I consider obsolete many aspects of a
traditional OS, as for the filesystem.
> As for not relying on a radical new approach,
> what is the advantage of
> (1) [free-standing lisp OS with traditional OS abstractions]
> vs (2) [a lisp implementation running on top of a traditional OS] ?
> If you're satisfied with the current OS abstractions,
> then (2) provides interoperation
> with another running "industry standard" OS,
> and provides you with free low-level drivers,
> without taking you anything (or does it?).
> If you're not satisfied with current OS abstractions,
> then what do you propose instead,
> and where/how does it conflict with current OS design/implementations?
> Before to look for a solution, we oughto identify the problem.
> What exactly does the lispOS consist in?
> Is it a set of Lisp primitives
> that cover all that is needed to program anything on a system?
> What model does it propose?
> Can people still implement their own GC?
> What if the system-provided GC is too slow,
> or lacks a feature essential to someone?
> Perhaps we've been too much discussing about implementation topics,
> and not enough what design exactly we're going to implement.
> In fact, many among you have already used lisp machines,
> and may find it obvious what you wanna do,
> while I, like most computer programmers on earth,
> including almost all the younger generation,
> have never had the chance to see/use a Lisp Machine,
> having be stupefied by other, `strongly hyped' languages and OSes
> (kudos to hbaker for the term).
> I appreciate any *affordable* pointer you could send to me;
> [are there running LispM's in France?].
Good question ... Symbolics never had that many systems in France,
I believe. They used to use them at France Telecom, a few here and
there for Graphics, one at Architecture University in Toulouse ...
> Kisses from the french Riviera,
> == Fare' -- firstname.lastname@example.org -- Franc,ois-Rene' Rideau -- DDa(.ng-Vu~ Ba^n ==
> Join the TUNES project for a computing system based on computing freedom !
> TUNES is a Useful, Not Expedient System
> URL: "http://www.eleves.ens.fr:8080/home/rideau/Tunes/"
Luca Pisati Voice: (310) 577-0518
Nichimen Graphics Fax: (310) 577-0577
12555 W. Jefferson #285 EMail: email@example.com
Los Angeles, CA 90066 Web: http://www.nichimen.com