[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 Jul 1, 2012, at 5:26 PM, Eli Barzilay wrote:

> 40 minutes ago, John Cowan wrote:
>> Form is content: the text of the draft is the only full expression
>> of the semantics that exists.  If I had my druthers, the formal
>> semantics would go away altogether, but then I'm prejudiced against
>> formal semantics: see
>> http://lambda-the-ultimate.org/node/4478#comment-69764 if you are
>> interested in my views.
> Coming from someone involved in designing a language, this would be a
> pretty bad thing to say.  When applied to Scheme, it becomes
> horrifyingly bad.
> (Yes, this sounds like flaming, but I personally think that this is
> *so* bad that it shouldn't go unhighlighted.)

John and I are here to horrify you!

You are just making a claim here.

Much as I love formalism, even when it's gratuitous, I have always felt uncomfortable with this particular denotational semantics. If the report said exactly *why* it's being provided that would help. Is it supposed to convince the reader of consistency or some other formal property? (Denotational semanticses provide models and proving consistency is usually the reason one creates a model.) Is it supposed to help implementors? It does not seem to be normative, and it is both incomplete and overdetermined (it describes an "extended subset"), so its role is mysterious. I'm not saying it's bad or useless, just that the nature of its potential utility is not made clear by the report.

If it gave an account of parameterize, guard, and dynamic-wind it might at least help answer the most vexing semantic puzzles posed by this draft. Without that it's sort of a tease.


> -- 
>          ((lambda (x) (x x)) (lambda (x) (x x)))          Eli Barzilay:
>                    http://barzilay.org/                   Maze is Life!
> _______________________________________________
> Scheme-reports mailing list
> Scheme-reports@x
> http://lists.scheme-reports.org/cgi-bin/mailman/listinfo/scheme-reports

Scheme-reports mailing list