Scheme language standardization process
We should fix this list's moderation policies. Gerald's reply that didn't seem to make it to the list:
--- Begin Message ---
- To: Andy Wingo <wingo@x>
- Subject: [Scheme-reports] Fwd: R7RS
- From: gjs@x
- Date: Mon, 29 Apr 2013 19:21:24 -0400
- Cc: "Arthur A. Gleckler" <scheme@x>, scheme-reports <scheme-reports@x>, gjs@x
- In-reply-to: <87obcxasgo.fsf@pobox.com> (wingo@pobox.com)
Hi, I hope you are enjoying Guile! Making a procedure have an invariant identity does not rule out the kind of procedure-integration optimization that you suggest. It is truly a bit more work for the compiler. Indeed, the compiler may integrate a procedure passed as an argument or returned as a value, but its identity need not change. This takes about 1 extra word of memory and a few extra instruction executions in the worst case. (And cycles are really free, compared with memory references, these days.) However, maintaining logical consistency is an important goal in language design. The main goal of specifying a language is not to make the job of the compiler writer easy. If that was the goal we could adopt "C". We should make a language we like. The compiler writers I know are surely good enough to handle small complexities like this. GJS
--- End Message ---
-- http://wingolog.org/
_______________________________________________ Scheme-reports mailing list Scheme-reports@x http://lists.scheme-reports.org/cgi-bin/mailman/listinfo/scheme-reports