[fitsbits] polar coordinates

[William Thompson]

On 12/11/2013 2:41 PM, David Berry wrote:
> On 11 December 2013 19:13, William Thompson <William.T.Thompson at nasa.gov> wrote:
>> We have a similar situation in solar astronomy.  One commonly used
>> coordinate system expresses coordinates in elongation angle from Sun
>> center, and the position angle counter-clockwise from the solar north
>> pole.  To work this into the WCS, I had to map the elongation angle E
>> into a latitude defined as E-90, i.e. make Sun-center the south pole of
>> the system.  I chose the south pole instead of the north pole, so that
>> the angle still increased in the usual fashion.  It's not an optimal
>> solution, but it does make it WCS-compatible.
> How do you encode the shift to put latitude zero at the south pole?

We don't.  The user has to make that conversion from latitude to 
co-latitude separately.

>> The part about taking the log of the polar angle should be fairly easy
>> to incorporate into WCS.  This is just a new spherical projection, which
>> would have to be properly defined and signaled with a unique projection
>> code.  There is a precedent to adding new projections which are
>> otherwise compatible with the WCS formalism.
> "Easy" is not a word often associated with trying to get changes to
> the FITS-WCS standard agreed and implemented! And I'm not sure having
> a separate projection code for every possibly interesting variant is a
> good idea.
>
> David