proposals for R5RS (sorry)

[Matthias Blume <blume@cs.Princeton.EDU>]

The discussion about how to guard ellipsis if they are not intended to
be ellipsis (in this round) seems to suggest using a binding construct
for the ellipsis.
	How about (sorry, I'm probably not up-to-date with the R5RS
syntax specification, I use the R4RS appendix):

<transformer spec> ->
	(syntax-rules (<identifier>*) <identifier> <syntax rule>*)

where the second <identifier> introduces the ellipsis for this set of
<syntax rule>s.  Would this cause any major problems?  We could even
make the second identifier optional using ``...'' as a default in
order to be backward compatible.

-Matthias