On Mon, Oct 31, 2011 at 13:38, Raymond Toy <toy.raymond at gmail.com> wrote:
> I also notice that ratepsilon is still 2e-8, even though maxima has been
> using double precision for years (decades?) now.
>
Argh! I thought we'd agreed multiple times (starting at least as long ago
as 2005) that ratepsilon should be much smaller, say 2.0e-15 (about 9
ulps). This is plenty of slop for returning "nice" rationals in the face
of rounding error.
-s