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

Re: [Scheme-reports] Technical question

On Thursday, Alex Shinn wrote:
> On Thu, May 26, 2011 at 9:59 PM, Aaron W. Hsu <arcfide@x> wrote:
> >
> > You ask for the collective WG1 voice on this, and unfortunately, I do not
> > think that we have a mechanism for this. However, we can file a ticket
> > concerning this issue and decide on it. Since even Chibi returns the
> > symbol ok for the above now (after Alex's fix), I don't think the above
> > code is particularly contentious anymore.
> I think people are looking for the exact wording in R5RS
> that describes this:
>      The escape procedure accepts the same number of arguments as the
>      continuation to the original call to
>      call-with-current-continuation.  Except for continuations created
>      by the `call-with-values' procedure, all continuations take
>      exactly one value.  The effect of passing no value or more than
>      one value to continuations that were not created by
>      call-with-values is unspecified.
> That seems unambiguous enough to me.
> I don't see why a formal WG1 interpretation of this
> is needed, unless someone is proposing that we
> change the wording.

Thanks for the clarification (from other wg people too).

> Regardless, this is a very far removed tangent from
> the original point, whatever that may have been.

I asked this question because John Cowan's post indicated that he
understands R5RS as saying that a conforming implementation can throw
an error instead of returning 'ok.  (This was indicated by his reply
when I claimed that an error means that Chibi's ex-implementation is
broken.)  Having a co-editor make such a fundamental assertion worries
me: if the answer to my question was "no" was indicated, then I'd view
it as wrong enough to (a) clarify the wording, and (b) make it known
that it is intended to make such implementations valid.  It is a major
conceptual change that should not be made without telling people about
it explicitly.

Since the question is "yes", I'm not worried about that.  (OTOH, I'm
very worried that a co-editor is making such mistakes, but I'll leave
this for later.)

Disclaimer: Yes, technically, this is a case of "someone is WRONG on
the internet".  But that joke is usually directed at useless
flamewars, with an outcome that is bound to be insignificant.  This
case is extremely different: the person in question is someone who is
in a position to change (or strongly influence) the language based on
the wrong understanding.  Ignoring mistakes is something that anyone
who cares about the language must not do.

          ((lambda (x) (x x)) (lambda (x) (x x)))          Eli Barzilay:
                    http://barzilay.org/                   Maze is Life!

Scheme-reports mailing list