Why such big divergence?

[eddyem]

Did you investigate why such a huge difference in coordinates between different libraries obtained? In this paper: https://www.researchgate.net/publication/326237049_A_comparison_of_SOFA_and_NOVAS_astrometric_software_libraries - the error between libraries estimated as 5-6mas. Not arcminutes...

[d33psky]

Hi eddyem,
That is a nice small error range that I had hoped to see but was not able to achieve.
Which is also why I put this simple code online as I found no other comparisons of these libraries.
Maybe I made errors with for instance things like TAI, TT, dut1 ? The test code is small, spotting errors should be simple ...
I've requested access to the paper you pointed to, waiting for access.
-- Hans