LispOS vs TUNES
Wed, 7 May 1997 02:25:41 +0200 (MET DST)
Dear Tunes and Tunes-lll subscribees,
for all those of you interested in coding a Dynamic OS *now*,
I believe that the LispOS project is a great opportunity:
* LispOS guys are going to implement a free OO OS based on CLOS
(Common LISP Object System -- the most elaborate Object System available)
* LispOS should steal all the good ideas from Lisp Machines
(a wonderful technology crippled by policies of over-expensive
proprietary hardware and software).
* It would provide orthogonal persistence with a Persistent Object Store;
I have proposed the tunes-lll mailing list to discuss
this implementation aspect.
* It would use the free CMUCL implementation of ANSI CommonLISP
(a compiler, that, on some architectures, beat GCC on many benchmarks)
* It would hack Linux, BSD, and/or HURD kernels where applicable
to get reasonable performance at memory management.
* It would participate in the free-CLIM implementation of CLIM standard
of OO Graphical UI
* A branch of the project is studying a portable VM
as a competition to the JVM that would be much more efficient for LISP,
as well as other languages.
* Commercial CL vendors are already interested to adapt their products
to LispOS if it succeeds.
* Low-level people are invited to implement the POS (above),
and/or develop a kernel on which to efficiently refound the system,
perhaps using the Flux OS-kit to steal Linux and BSD drivers.
All This does not mean that TUNES is a dead project. Not at all.
TUNES and LispOS have different goals.
LispOS starts from a technology that already existed, but was proprietary,
and strives to provide a free implementation with enhancements
(much like Linux vs UNIX). TUNES develops new concepts,
some of which are present in LispOS (orthogonal persistence),
some of which are not (Reflection).
I do believe that Reflection and Meta-Programming
are the future of Computing in general, and Operating Systems in particular.
Only they are not ready technology yet,
and that'd be a waste that everyone wait for them before to begin coding.
I'll be working on making them (freely) available,
and will appreciate any help you can give me about that,
together with any idea (or code) you have about software features
that an innovative system ought to support.
Meanwhile, I invite you code in CLOS an integrated OO environment,
which LispOS offers us to develop *now*.
When I have a Reflective Language with Invariants,
I'll use it to reimplement LispOS on top of a reflective kernel,
and use LispOS as the basis for future TUNES developments;
being high-level, LISP should interact correctly or even nicely
with the TUNES HLL, whatever it will be.
Well, whatever you do, good luck!
== 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