application framework

[zilla@overdrive.nj.nec.com (John Lewis)]

Without knowing anything about either macapp or bedrock, I nevertheless
say that requiring an MCL application framework to emulate either
of these is a mistake!

C++ is object oriented, but it is not a flexible or high level language
in other respects.  The compromises that are made in coding for C++
could cripple a lisp application framework.  I think the designers
of the application framework should certainly pass the design past
people familiar with macapp/bedrock, and make the MCL framework conform
only where it does not hurt to do so.

Also, the issues of machine and language portablilty are quite different
and should be separated.  Lisp/clos is sufficiently high level that
i could imagine an application framework (and hopefully MCL itself)
being ported to other apple-supported architectures.  Meanwhile, 
for the academic/lisp community, we have CLIM and possibly other 
interface builders (Action?).
  Cross language portability is a wierd concept to me.  Does anyone really
want this?  I'd rather wait and hope that MCL itself (with the help
of this application framework) is sufficient to develop an application.
I don't think lisp will ever be suitable for writing an init, da, or
microsoft word.  It doesn't seem unreasonable that MCL could be used to
write some complicated/specialized applications (say, a limited market
CAD program) fairly soon however.

In summary, MCL's application framework development should go on 
in parallel to bedrock, but not be limited by it.  MCL does not
have the same size development team as macapp, but it also takes
less time to get things done in MCL.  
X amount of time from now, I expect that the smaller MCL development
community will have/continue to have an environment which is competitive
with and sometimes superior to whatever the C++ people are doing!

j.p.lewis/nec research