On Mon, 10 Oct 2005, Michael Sperber wrote:

>> An external representation would be tricky to construct for instances of
>> generative record types, but for non-generative types it should be
>> possible, and would certainly be useful.
>
>Yes, and I expect this will be an addition at some point in the
>future.  (Non-generativity is partly there for exactly that purpose.)
>For this SRFI, we decided to punt on this, though, because of the
>expected technical complexity.

I think there's a legitimate expectation and widespread implementation
of the idea that a scheme implementing simple records (aggregates)
ought to be able to read and write them in an external syntax.  And
while I'd like to see that _functionality_ documented, promoted, or
possibly standardized, I don't think there's any consensus yet for any
_particular_ external syntax.

			Bear