I considered the ((call/cc proc)) trick, but decided that it "required
cooperation from the client" -- you have to use that double-paren trick
at the call/cc site.

But you have proposed it as a sort of "different" kind of continuation,
which is just a different perspective on the same observation, but one
that makes it more clear to me that this is no worse than the "kontinuations"
I proposed in my message.

So, fine!

Do you think this should go in the SRFI, or be left out? Possibly discussed
in the rationale section, since it isn't necessary to define any new
primitives?
     -Olin