What I don't want in R^5RS

[kend@data.rain.com (Ken Dickey)]

Like many of you, I would like R^NRS to describe the core of the
Scheme language and not be a feature dumping ground.  On the other
hand, there are a number of interfaces which are difficult to
implement using core Scheme but which I would like to be standard
where they are implemented.

The classifier I use is that features which are language related (e.g.
multiple values) pertain to R^5RS and those which are primarily data
type oriented (e.g. string ports, binary i/o, bitmaps, menus, etc.)
are not.  The latter category are either models which are proposed and
made public as "library" code (e.g. window systems) and/or
documentation, or are "standard" features (e.g. string ports) which
should conform to a given interface.

So it appears to me that in order not to have `stuff' dumped in R^NRS,
there should be a mechanism to review non-R^NRS stuff.  The question I
have is, should this group be the medium of discussion and if not,
what, how, & who ?   

Just a thought [before net.lang.scheme boils over],
-Ken						    kend@data.rain.com