Lisp vs. Scheme
Tue, 24 Mar 1998 01:54:09 +0000
Gavin E. Gleason wrote:
> Aside from this CL has a huge code base of very complicated
> and powerful programs that are largely portable between systems. PLOB
> is a persistant object system already writen for CLOS. There are also
> a number of non-standard Object systems writen for CL that are by no
> means "toys". KR (for garnet) and LOOM come to mind. So even if you
> don't like CLOS you could find something in CL. CMUCL also has an
> emacs-clone that has already been writen, Hemlock. Seems to me you
> could save a heck of a lot of time just extending and integrating CMUCL.
> I'm by no means a CL-fanatic. I do lots of scheme programming
> (probably as much as in CL), but I think that a schemeOS project is
> doomed to reinvent lots of the stuff that is already present in lisp.
There is no need to reinvent it. All we have to do is steal it!
If any of us really cared about "huge code bases", we would be
hacking C under MS-Windows. If you are happy with these CL apps
that already exist, then I can't see why you care about LispOS.
Just run these apps on whatever OS you have available now.
IMHO, LispOS is all about what can be done by building lots of
little modules on top of each other and integrating, integrating,
integrating to remove the monolithic nature from the conventional
way of doing things. Instead of having one hugh MS word app,
there are lots of little modules that build upon one another.
While I'm sure there are a few CL bits of code lying around that
can fit in this paradigm, I'd say a lot of the code has a
different mind-set anyway. Those "very complicated and powerful"
programs for instance.
The limitations of Scheme and Lisp can be fixed, and would need
to be to get everything just right. But personally I think a
Scheme system would end up being better in the end. And it's not
clear that using Lisp would be a faster means to that end anyway.