This is the mail archive of the
guile@sourceware.cygnus.com
mailing list for the Guile project.
Re: New module system wishlist
- To: neil at ossau dot uklinux dot net
- Subject: Re: New module system wishlist
- From: Clark McGrew <mcgrew at ale dot physics dot sunysb dot edu>
- Date: Tue, 25 Jan 2000 10:57:14 +0900
- CC: clark dot mcgrew at sunysb dot edu, guile at sourceware dot cygnus dot com
- References: <E12CfSt-0004tJ-00@boxer.icrr.u-tokyo.ac.jp> <200001242147.VAA00422@ossau>
- Reply-to: clark dot mcgrew at sunysb dot edu
>>>>> "Neil" == Neil Jerram <neil@ossau.uklinux.net> writes:
Neil> I think what I had in mind, when I wrote the above, was that
Neil> we should not feel afraid of making the new module system
Neil> completely different in design from the existing one, if
Neil> that is what emerges from the current discussion of
Neil> desiderata. I then unnecessarily mapped "difference of
Neil> design" onto "difference of interface".
In which case I agree with you. I assume that "backward
compatibility" on the wish list means "superficial syntatic backward
compatibilty" or practically, "doesn't break old code".
Neil> Actually, there is still _an_ argument against such
Neil> emulation, which discounts the pain of having to change
Neil> existing code against the potential future confusion of
Neil> having a "non-orthogonal" module API.
For me, the ideal module system will include the superficial highlevel
syntax of the current system. I think the only "fixed points" are
"(define-module (module-name))", "(use-module (module-name))" and
"(export a-symbol)". I'm not even arguing that all of the current
features of those forms have to be supported.
Cheers,
Clark