[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: tail recursion

>   If you don't like
> the requirement for proper tail recursion, fine - argue for its
> removal - but I think that it is dubious to argue that we don't know
> what the requirement means.

Well, I'm not sure it's as dubious as all that.  I think there are
implementation strategies that we would all agree are tail recursive, and
others that we would all agree aren't, but I bet there are still others
that would start violent arguments on this mailing list.  Luddy's problem
is quite literally that he would like to experiment with new and quite
unconventional implementation strategies and he would like a rigorous
statement of where that boundary lies (presumably accompanied by a
convincing rationale) so he knows whether it makes sense to classify a
particular implementation as tail-recursive or not (and why).

We shouldn't assume from the foregoing discussion that we don't have any
idea what proper tail recursion means, but I don't think the discussion
demonstrates that we have a solid, unambiguous test for it either.  -Bert