Re: Issue: ELIMINATE-FORCED-CONSING (Version 3)

[masinter.pa@Xerox.COM]

We need to be careful about cc'ing the relevant parties on discussion on of
issues. I've forwarded to jrg those messages that originally went to
cl-cleanup only.

This issue was not distributed to X3J13 prior to the meeting, I think. I am
reluctant to have it on the list with its current name and am tempted, at
this late date, to rename it.
(SEQUENCE-FUNCTIONS-CONSING:ADD-TARGET-KEYWORDS) or some such.

Jim Allard, I believe, made the comment that, given a dynamic-extent
construct, it is possible to do "cons-free" programming with some larger
awkwardness by writing an idiom where the new sequence is generated with
dynamic extent and then the old sequence is either copied or modified.

I wonder whether some special purpose recognizition of idiomatic nesting of
REPLACE with a sequence function inside compilers or optimizers might well
have the same benefit without increasing the apparent complexity of the
language. (I believe some APL compilers work this way.)