John Cowan wrote:

> > As has been pointed out to me, the WG1 archives prove that a plurality of
> > the WG1 members who voted on this intended to require the new contour
> > semantics, but the R7RS (small) document fails to do that, and the R7RS
> > also fails to note any change from R6RS semantics in this regard.
> >
> Since this was an editorial failure, it should become an erratum.

Although the 5 members of WG1 who voted for this change might regard
it as an editorial failure, some of the 54 members of the community
at large who voted to ratify the R7RS (small) standard might regard
your proposed erratum as an attempt to change the substance of the
standard after it has been ratified and accepted.

During the ratification vote, most voters would not have known five
members of WG1 had voted to forbid the splicing semantics allowed
by the R5RS, required by the R6RS, and (accidentally) allowed by
the R7RS document itself.

Had voters realized WG1 intended to make this change, there would
have been at least one formal comment opposing the change.

Four years have passed.  Two implementations of the R7RS (Chibi and
Larceny) have been using the splicing semantics.

Larceny v1.3 will be the second implementation of R7RS Red Edition.
The current development version of Larceny, when run in -r7strict
mode on a recent Linux machine, passes all of Larceny's 8784 R7RS
tests.  In its -r7 and -r7r6 modes, designed for interoperability
with R6RS code, Larceny fails only the two tests I added yesterday
that require let-syntax and letrec-syntax to introduce new lexical
contours.  You can't accuse Larceny of frivolous non-conformance.

Will