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

Re: [Scheme-reports] Formal Comment: clarify the semantics of the dynamic features

On Fri, Jun 29, 2012 at 12:38 PM, Richard Kelsey <kelsey@x> wrote:
>   Date: Fri, 29 Jun 2012 11:24:36 -0700
>   From: Alex Shinn <alexshinn@x>
>   > Richard Kelsey scripsit:
>   >
>   >   Date: Thu, 28 Jun 2012 17:55:54 -0400
>   >   From: John Cowan <cowan@x>
>   >
>   > The report should describe the semantics of dynamic variables, not one
>   > particular implementation.
>   The report describes the semantics in terms of dynamic extent,
>   so the analogy to dynamic-wind seems appropriate.
> No, the description of dynamic bindings doesn't mention dynamic
> extent.  The current text uses the phrases "during the evaluation of
> the body" and "evaluated in a dynamic environment in which ..." which
> in the presence of call/cc and dynamic-wind, and in the absence of a
> definition of dynamic environments, are far from clear.

Ah, I very deliberately avoided the use of the term "dynamic
environment" in the original description - it seems to have
snuck in.  I'll review this later when I get a chance.

>   We do need to be careful here, and the WG has not yet had
>   time to fully review the non-call/cc-based dynamic-binding
>   research pointed out by Oleg.
> Can you give me a pointer to this?

It's in the thread starting here:


I'll try to summarize this on the wiki later.

>   >   Editorial tickets #427 and #428 created.  Ballot ticket #429 for new
>   >   formal semantics created.  If nobody steps up to do this and review it
>   >   before the last ballot, it will be closed.
>   >
>   > If the WG makes changes to the language that require changes to the
>   > formal semantics, then they need to change the formal semantics.  That
>   > seems like part of the job.
>   The formal semantics only covered a small subset of the
>   language to begin with.  The change in question is equivalent
>   to adding a formal semantics for dynamic-wind, which was
>   already missing from R5RS.  I agree, of course, that we should
>   fix this, or consider using (a subset of) the operational semantics
>   from R6RS (which again is a subset of the language but at least
>   has dynamic-wind).
> My comment was meant to be a more general response to John Cowan's
> planning to close a ticket if there was no one on the committee
> willing to work on the semantics.  I was actually thinking of eqv?,
> where the report and the formal semantics no longer agree.  In that
> case something does need to be done.

The definition of eqv? is still under debate, but if the
end result is incompatible with the formal semantics
we will definitely update it.


Scheme-reports mailing list