> without identifier syntax, at least
> what looks like a variable *is* a variable and not some arbitrary
> expression under the table.

177 will work with or without identifier syntax, as long as the keyword
lambda object from the identifier-syntax version can be passed to 177's
call/kw and still work somehow.

> I am very reluctant to deal with a notion of "most relevant systems".  Very
> early in this process I said that R4RS was obsolete and was pulled up very
> sharply by more senior Schemers.  I *am* willing to say that SRFI n depends
> on some other SRFI or some RnRS feature.

I would also root for a very broad definition of relevant Schemes. If we
serve only the most popular ones, we are in effect advocating that they
stay that way. The implementors of currently unpopular Schemes often
appreciate when we include their implementation in SRFI/RnRS
considerations, and that kind of thing motivates them to improve it more
actively. Small or unusual Schemes are also a great place for innovation
in the language, since they don't have as complex internals and so much
legacy burden they can be more agile.

So I would rather advocate that each library or feature have the
smallest "reasonable" set of dependencies (obviously this is a somewhat
fuzzy notion, but we shouldn't introduce gratuitous incompatibility or
complexity). I regret the loss of a syntax-rules implementation for 177,
but I think in this case the argument for syntax-case/ER is very
powerful, and also define-macro or a special form is very simple to
implement even for small Schemes.