> The word "formal" in "formal semantics" is not important.  The "formal
> semantics" are just the, well, formal, description of the semantics of
> the Scheme language as defined in the earlier chapters of the R7RS.
> Thus it suffices and is probably less confusing if we just talk about
> "semantics".

Fair enough.

> The root problem is that the arity of a procedure has no well-defined
> meaning in the semantics of Scheme.

It seems you're right. That explains a lot!

> mailing list of SRFI 102: "Finally, I just want to say that I do not
> think procedure arity inspection should be incorporated into future
> Scheme standards without a much more convincing rationale than this SRFI
> provides."

That may be the right call for RnRS. Nevertheless, a SRFI outside of
RnRS would almost certainly be useful in its own right.

> So, please excuse my role as a devil's advocate in this discussion, but
> the issues that were already present with SRFI 102 should be solved
> before SRFI 191 becomes more than a SRFI, namely part of some standard
> like R7RS-large.

No problem. I appreciate your attention to detail and in this case as
well as others your objection is perfectly reasonable.