On 3/28/07, Abdulaziz Ghuloum <aghuloum_at_cs.indiana.edu> wrote:
.
>
> Maybe I didn't understand what you were suggesting. Are you
> proposing any specific canonical semantics of letrec that does
> not involve side effects but is still useful[*] for Scheme?
No. And I realize this is wimpy, but it is just a suggestion, not
a formal proposal.
I just think the canonical semantics for LETREC should not
require SET! and that a fixed-point operation should be an
allowable alternative. This would imply that programs that
could observe the difference between these two sorts of
implementations would be considered `in error'.
I don't *think* this will outlaw too many `useful' programs,
but I could be wrong.
--
~jrm
Received on Wed Mar 28 2007 - 18:25:28 UTC