branch cut fixup for atanh(bfloat(x))



On 2/27/06, Richard Fateman <fateman at cs.berkeley.edu> wrote:

> I guess my (perhaps naive) view of this is that
> the symbolic simplifications we make may depend on
> the branch cut definitions as well, and the numeric
> values, while important, are incidental to the
> branch cut problem.

As it happens, at the moment the symbolic and numerical results agree
about branch cuts. I personally believe that we should keep it that way.

By the way, I made them agree by changing the numerical branch cuts.
I made no other changes.

For what it's worth,
Robert