I'm announcing the first release of my open source project named
"PreciseFloating" hosted on sourceforge at
http://sourceforge.net/projects/precisefloating. This version includes
implementations for the round-to-negative infinity and
round-to-positive infinity directions of rounding mode and a lot of
documentation. Also, simple continued fractions arithmetic is
included, and actually used to compute the exponential function with
directed rounding. As I'm not aware of any similar proofs, I included
my proof the convergence of the linear/bilinear transformation
algorithms in the docs/docs\continuedfractionsarith.pdf paper. Anyway,
I'm willing to learn about any similar work.