[plt-scheme] Please help test version 359.100

From: Robby Findler (robby at cs.uchicago.edu)
Date: Tue Nov 14 11:35:45 EST 2006

At Tue, 14 Nov 2006 11:23:58 -0500, Matthias Felleisen wrote:
> On Nov 14, 2006, at 11:16 AM, Robby Findler wrote:
> > At Tue, 14 Nov 2006 11:12:51 -0500, Dave Herman wrote:
> >>>>    Since the value returned by "for-each" is left unspecified
> >>>
> >>> But it isn't left unspecified. It is specified to be the specified
> >>> thing that the r5 report just calls "unspecified". (how confusing is
> >>> that?!)
> >>
> >> I had thought it was only in R6RS where they explicitly specified a
> >> thing called the "unspecified value," whereas in R5RS it's left  
> >> ambiguous.
> >>
> >> I did a quick grep of R5RS and found that the phrase "an unspecified
> >> value" occurs 7 times in R5RS (all in section 6), whereas there  
> >> are no
> >> occurrences of "the unspecified value." The use of the indefinite
> >> article "an" suggests that for *each* occurrence of the phrase, the
> >> value in question may be any value.
> >
> > I'm not sure how that refutes what I wrote (was it meant to?)
> >
> > I think the relevant parts of r5rs were quoted earlier in this thread
> > by Jacob and me. No grepping required.
> I think it refutes it unless I misunderstood what you and Jacob are
> arguing. According to your emails, a Scheme semantics uses a single
> value dubbed "the-unspecified-thing" and injects it into the execution
> wherever the semantics says "unspecified." For all we know, you could
> use 42 at all these places.
> Dave and I argue that an implementation can make up a new random value
> -- as long as it is a Scheme value -- and use it wherever the semantics 
> says unspecified.
> If you tried to say the same thing, your English comprehension is
> very different from mine.

Huh? I don't see how English comprehension is esp. relevant here. We're
talking r5 comprehension, an entirely different beast.

I think Jacob and I only said it cannot be *multiple* values, unless
your system implicitly converts multiple values to a single value.

No one has disagreed with that.


Posted on the users mailing list.