different alternatives for lispOS

[Luca Pisati]

Christopher J. Vogt wrote:
> 
> At 2:01 PM -0500 4/25/97, Kelly Murray wrote:
> >> What I want is Genera that runs on mac or pc hardware, commodity stuff.  I
> >
> >Might work ok, but I don't see it happening.  The new Symbolics can hardly
> >get it to work on Alphas (maybe they just need a lot more help?)
> 
> I Don't know how many people are working at Symbolics these days, but I'll
> bet it isn't many.  There are probably more former Symbolics employees at
> Harlequin than there are current employees at Symbolics.
> 
> >It also really *is* old technology (e.g Flavors)
> >and is a very complicated system, full of historical baggage.
> >Emulation will always perform poorly relative to native complilation,
> >and it has little chance of being available on other platforms given
> >the struggle to put it on just one(*).
> >It also doesn't have real persistent objects.
> >Statice is essentially a rdb with an o-o interface.
> 
> While it's true that Genera was originally built on Flavors, it supported
> CLOS by the time I left Symbolics circa 1989.

It supported "some kind" of CLOS, it didn't support any MOP,
and I do not think it was really optimized at all.

> There were tools to convert
> flavors code to CLOS code (I don't know how much of the system was
> converted), and CLOS isn't *THAT* different from flavors.

Yes, but converting from Flavors with generic-function is pretty simple,
converting from Flavors with message-passing can be a little bit more
complex.

CLOS is not that different from Flavors, but programming cleanly with
CLOS it is very different.  Whoever programmed with Flavors tended to
abund in PROPERTY-LIST-MIXIN and MIXINs in general.  Those things are
easily ported to CLOS (as you know), but can create big structural
problems, in part because the topological sorting algorithm used by CLOS
is very different than the one used by Flavors.

> Further, I don't
> know that I was specificly thinking of using actual Genera code, I was more
> thinking of it as a functional model.  I certainly wouldn't want anything
> that isn't native compiled.

Yes, I do agree on this, since big protions of Genera had great models
behind.

> [...]
> 
> Christopher (Chris) J. Vogt
> mailto:vogt@novia.net
> Omaha, NE
> http://www.novia.net/~vogt/

--
Luca Pisati		       Voice:	 (310) 577-0518
Nichimen Graphics	       Fax:	 (310) 577-0577
12555 W. Jefferson #285        EMail:    pisati@nichimen.com
Los Angeles, CA 90066          Web:      http://www.nichimen.com