[R6RS] smaller condition hierarchy

[William D Clinger]

Matthew proposed a smaller condition hierarchy, along
with background on how MzScheme's hierarchy evolved.
Thank you, Matthew.

As I understand it, MzScheme's exception system does
not support continuable exceptions, or else continuable
exceptions are used but rarely in MzScheme.  I believe
a more detailed hierarchy becomes useful only when
exceptions are continuable.

So far as I recall, the &no-infinities and &no-nans
conditions are the only ones so far that correspond
to exceptions that definitely need to be continuable,
but we may want to require some of the others to be
continuable as well.  That's just something to think
about when voting.

I suspect that a relatively small condition hierarchy
is appropriate for the R6RS, but a compatible and much
more detailed hierarchy might also be proposed in some
SRFI.  The votes we cast might influence that SRFI, so
it might be worth thinking about the detailed hierarchy
even if you would advocate a small one.

Will