good science and lisp->C case histories

[Kenneth Forbus <forbus@A.CS.UIUC.EDU>]

   ....Certainly one hears on the street
   that such translation is being done all the time (there are only two
   kinds of lisp applications, those already translated into C and those
   that are going to be, etc).

Jeepers.  I've never written a line of C code in my life.  And I never
intend to.  Maybe I don't write applications??

Lookit, the differences have to lie in (a) what environmental services
are provided and (b) what the libraries are like.  It is tough to
control for (b).  I think it's a pretty sure bet for (a) that if you
make programmers do explicit storage reclaimation, when they can do it
correctly it will be faster than letting them rely on garbage
collection.  If your job is to make some system run like blazes on a
micro, I'm sure you may want to force programmers to do such horrible
things.  People sometimes resort to such tricks in lisp for
performance, too.

The question is, what sorts of tradeoffs do you want a language to
support?  Lisp supports rapid prototyping better than anything else, I
think.  If circumstances become such that one wants to squeeze the
last drop of performance out of a particular piece of iron, one may
want to switch to a software environment that is optimized for doing
that (i.e., C).