[R6RS] incomplete draft of revised i/o

[William D Clinger]

Mike wrote:
> Another question: Why is eqv? on transcoders specified?  I can see how
> a specific implementation might exploit it, but am unsure how it would
> be useful for programs.  (I know this was the case even before the
> revision, but just caught my eye just now.)

As we have seen, some people think they want to use eqv?
to compare arbitrary records or NaNs.  Comparing transcoders
seems more useful to me than comparing arbitrary records or
NaNs, and I was trying to remain compatible with the general
drift of the R6RS toward specifying things that would result
in non-portable code if left unspecified.

If the editors think the behavior of eqv? should be left
unspecified on transcoders, then I think we should add a
binary-transcoder? predicate to make it easier to recognize
the binary transcoder.

Will