Embedding VMs

[Michael Korns]

Marc,

I understand your point about wanting to "Do it right" and not bring along
useless baggage. 

This is a noble goal, and one which I'm sure you will have luck with in the
future. However, as I've already mentioned. Any effort which is not Web
enabled, or which excludes the Java supporters, etc. would not be of
interest to my company.

----------
> From: Marc Wachowitz <mw@ipx2.rz.uni-mannheim.de>
> To: lispos@math.gatech.edu
> Subject: Embedding VMs
> Date: Sunday, April 27, 1997 11:19 AM
> 
> The idea to create e.g. a direct superset of the JVM is misguided and
would
> only create useless baggage. Merely because almost everyone gives out
some
> marketing hype about being a true FOO-supporter (FOO being Java, ActiveX,
> CORBA, or whatever may be the buzzword of the day), this doesn't mean
that
> we need to let all that mess "design" our system. Let's remember that
Lisp
> is about easy and powerful embedding - not about hard-wired emulation.
Just
> add a translator from whatever VM you wish to support to Lisp - e.g. a
real
> compiler from JVM to Common Lisp would be merely tedious work, assuming
some
> suitable foundations (e.g. you probably want to have direct IEEE floating
> point arithmetic implemented, since software emulation would be too slow
for
> many cases), which are mostly good to have for a modern Lisp system
anyway.
> 
> -- Marc Wachowitz <mw@ipx2.rz.uni-mannheim.de>
>