Thanks for your considered reply.
Though most people prefer lowercase for legibility, I don't feel strongly
about which case is used. The problem ACL has with the current defsystem
code is that in some places objects are referred to by uppercase "(defun
FIND-SYSTEM..." and in other places by lowercase "(find-system..."). As
long as the case is consistent, Franz's "kludy solution" will work.
I'm fairly new to Lisp. For my own code, I also use keywords in
defpackage. As to minimally change the defsystem code and keep the
defpackage macro invocation consistent the rest of the code, I left the
terms in defpackage macro as strings.
I'm not very familiar with Lisp implemtations other than ACL. Should the
defpackage expressions in the non-ACL sections also use keywords or
uninterned symbols rather than strings?
> I am reluctant to include these patches in the DEFSYSTEM distribution.
> Using Franz's case sensitive images is non portable and the workaround
> that you need to use to accomodate such images is (IMHO) aesthetically
> ugly (i.e. the use of uninterned symbols or keywords) and it also
> works well only in ACL, AFAIK.
> Having said so, your changes are not quite correct.
> > From: Kevin Rosenberg <kevin@...>
> > ;;; For CLtL2 compatible lisps...
> > #+(and :excl :allegro-v4.0 :cltl2)
> > -(defpackage "MAKE" (:nicknames "MK") (:use "COMMON-LISP")
> > +(defpackage "make" (:nicknames "mk") (:use "common-lisp")
> The correction should be
> (defpackage :make (:nicknames :mk) (:use :common-lisp) ...
> (Other changes should be done consequently.
> Let me say that I think it is a very unfortunate state of affairs that
> Common Lisp is case insensitive and uppercasing. But Franz's solution
> is, IMHO, kludgy, non portable and, above all, not backward compatible
> with the ANSI standard.
> Apart from this, I will make the necessary changes to the
> distribution if people really feel that they really really need to use
> Franz's case features.