Comment 2 for bug 965811

Thanks for the explanation!

After setting pressure to zero I am finding a difference (separation1 - separation2) of 23 arcsec for the one example I gave above.
Is the accuracy of the trigonometry calculations really this bad?
Or is there another real effect?

I am trying to use pyephem to test results from another coordinate package, for that I have to be sure the pyephem results are correct approximately at the arcsec level.

Thank you for your help!