RE: Beam slopes in PMX

[Simons, Don]

Joel Hunsberger wrote

>would it help (restore
>the once manifest behavior) to change the test so as to control the
>precision of the result?  Such as
>
>if (abs(slope2) - abs(slope1) + precision_limit ) .lt. 0 ) slope=slope2

That will be part of the solution, but not the hard part.  It would
unambiguously single out the cases where the two absolute values are
effectively equal.  But it won't tell you which of the two competing values,
or what other value, to use.

--Don