[r6rs-discuss] Re: [Formal] Allow compilers to reject obvious violations

[William D Clinger]

I am posting this as an individual member of the Scheme
community.  I am not speaking for the R6RS editors, and
this message should not be confused with the editors'
eventual formal response.

Matthias Felleisen quoting me:
> > Why, by the way, do you regard the requirements of a
> > "formal and executable" specification as less whimsical
> > than the requirements of an informal specification?
>
> 1. This is a truly strange remark from the editor of three reports
> that attempted to specify the meaning with a denotational semantics.

Not nearly so strange as the rest of your remark 1:

> My own research has suggested time and again that denotational
> semantics isn't truly up to the task in many cases so I prefer
> operational semantics, like SML's. The one that comes with R6RS has
> the advantage that it is a semi-algorithm.

Reading that, a less charitable reader than I might
conclude you do not realize that both denotational and
operational semantics can be formal, that both can be
executable, and that neither has to be executable.

People are discussing real issues in this thread.
Those issues have little to do with formal semantics
or formal systems.  IMO it would be silly for us to
continue to discuss the limits of our understanding
of formal semantics and systems.

Will