make PolarizationState more flexible

[pdowler]

ObsCore and FITS WCS allow a limited set of states

actual data files often use combinations that better describe the polarization axis (even if degenerate aka 1 pixel)

CAOM-2.4 allows an extended set of values from https://listmgr.nrao.edu/pipermail/fitswcs/2008-March/thread.html

Recent metadata requests include additional things like Q + iU...

[pdowler]

The only viable solution appears to be to define a vocabulary that can evolve over time and define Plane.polarization.states as a set of those vocabulary terms.

[pahjbo]

It is an irony that the FITS polarisation codes were rather dominated by what was done in radio interferometers - and yet there are still missing states - but I think the real conclusion of that thread is that it is rather difficult to come up with a vocabulary that covers the possible cases - I think that this https://listmgr.nrao.edu/pipermail/fitswcs/2008-March/000425.html contains that insight - however it is not an easy problem to solve as no-one has done it yet - it should be solved upstream of CAOM though....

[pdowler]

I agree that it should be solved upstream. From the CAOM point of view, I want to refer to an external resource and for technical reasons that is best served by a vocabulary. The IVOA vocab could be the place where the community discussion actually comes together, or it could be more like the UAT where IVOA simply endorses/integrates something from elsewhere.