Re: requiring proper tail recursion

[ramsdell@linus.mitre.org (John D. Ramsdell)]

I think Will's version of the tail-recursion requirement is very
good.  I agree with every point he made.

William D Clinger <will@ccs.neu.edu> writes:

> I would prefer to keep the discussion of tail-recursion in
> Section 1.1 brief, because that section is after all a brief
> overview of many different things, and just add "See section
> X.Y." as a pointer to a fuller discussion.

Good idea.

>   1.  For a COND clause of the form (E0 => E1), the implied call
>       is in a tail context iff the entire COND expression is.

An oversight on my part.  Good fix.

>   3.  EVAL is not required to evaluate its argument within a
>       tail context; see below.

I just copied Richard's text, I'll let him debate this.

>   2.  Top-level expressions are not counted as tail contexts.

On this point, I think I made a too literal a translation of the
definition of tail combinations using contexts into Scheme.  Opps...

Good show,

John