Hi Marc,

we have reviewed the issues surrounding SRFI 0 resulting from the
discussion so far.  We've also studied a range of Scheme
implementations for compatibility with the ideas of SRFI 0.  The
current draft covers a lot of the ground, but not all of it.

In particular, some Scheme implementations take the stance that no
features beyond R5RS are present "by default"; any additional features
require explicit imports.  It would be great if we could have a SRFI 0
which would cater to this situation, without impeding its viability on
more traditional systems.  This has turned out to be a complicated
task; the mailing list discussion had already uncovered that there are
non-trivial scoping issues to be worked out.

Since the mailing list discussion has already shown most of the issues
at hand, we decided to just go ahead and suggest a concrete revision
of SRFI 0 which caters to these requirements.  It is implementable
with essentially the same machinery as the current SRFI 0 draft, but
allows for a wider range of implementation strategies, and thus maps
more naturally to the various native language extension mechanisms of
current Scheme systems.  Functionally, the revision suggestion
subsumes the current draft.

We hope you like it.  Let us know!

Dave, Mike, and Shriram (The SRFI Editors)