Andre van Tonder <xxxxxx@later.het.brown.edu> writes:

>    I don't know, but a related question:  Does the spec in principle
>    disallow extensions for exact reals that are not exact rationals,
>    in the sense of Edalat, Potts, etc.?
>    (See e.g. http://www.doc.ic.ac.uk/~ae/exact-computation/ ).

If you're asking whether the spec disallows this for the operations
specified in the section on exact arithmetic, then you're right---they
only work on exact rationals.  This is debatable, I guess, but one of
the purposes of that section is to have completely reproducible
arithmetic.  You lose that if you make these operations open-ended.

--
Cheers =8-} Mike
Friede, Völkerverständigung und überhaupt blabla