> It's still polite to ask them if they want the patch. Plugging in a PDA
> fortran routine would never work for that.
No it wouldn't. I only later saw that the NR-derived code was in C.
My point in mentioning it was that there are bound to be plenty of
implementations of Gauss-Jordan elimination out there. In Starlink we
largely removed licensed code when we replaced all the NAG-library calls
with the PDA library we had assembled some 20 years ago. The question
of whether ESO would modify rtd/skycat in the same fashion is a
different matter.
Malcolm
|