ObsCore pol_states ordering

[Patrick Dowler]

Hi all,

I have an issue trying to reconcile the order of po_states values as
specified in ObsCore with what I'm doing in CAOM.

In ObsCore-1.1 B.6.5.1 the order is specified as

I, Q, U, V, RR, LL, RL, LR, XX, YY, XY, YX, POLI, POLA

The first N-2 values are from FITS WCS Paper I (table 7) and that seems to
define the canonical order (not the "coordinate value" order since the
circular and cross coordinate values are actually negative). The latest
FITS standard (
https://fits.gsfc.nasa.gov/standard40/fits_standard40aa-le.pdf) (Table 29)
includes the same set of values as paper 1.

Issue #1: POLI and POLA are not defined, they are in the 2010 polarization
note by Richards & Bonnarel

In CAOM, we use an extended set of values that includes POLI and POLA (but
also others). I got the names and coordinate values from a FITS WCS
discussion (fitswcs mailing list, 2008-March, message 000408 -- can't find
a working link). In my implementation, I just sandwiched these extended
values in between V and RR because the coordinate values are positive and
follow after V (see http://www.opencadc.org/caom2/#PolarizationState for
model, see
https://github.com/opencadc/caom2/blob/master/caom2/src/main/java/ca/nrc/cadc/caom2/PolarizationState.java
for mapping to numeric coordinate values)

Issue #2: B.6.5.1 says the ObsCore list/order is "compatible with the FITS
list" but  I can find nothing concrete in FITS to base this on. It is
inconsistent with what I'm doing in CAOM.

I can (probably harmlessly) change the order in my java code to match the
declared order in the model and that would match ObsCore, but I'd like to
make sure it is actually "right". Since CAOM has more values that ObsCore
(and they are interleaved a little) I am being cautious.

Is there a missing reference from ObsCore to something concrete?

Possible errata for above issues?

--
Patrick Dowler
Canadian Astronomy Data Centre
Victoria, BC, Canada
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.ivoa.net/pipermail/dm/attachments/20200617/9161bdc2/attachment.html>