[fitswcs] TAN+poly & astrometric discussions in WCS paper-2

[Mark Calabretta]

On Tue 2000/01/11 09:34:00 -0000, Peter Bunclark wrote
in a message to: Don Wells <dwells at NRAO.EDU>
and copied to: fitswcs at NRAO.EDU

>Your client is not allowed to access the requested object.'' 

Now fixed.

>Since I really wanted to get this whole thing finalised, I didn't bring
>up the following at ADASS; however, there seems to be enough loose ends
>that another might as well be added; the point I want to make is that
>Schmidt cameras are concentric optical systems to quite high order
>(Seidel perfect) and hence are fundamentally ARC-projection systems;
>by starting with TAN you have to have a higher-power, and hence less
>stable, polynomial correction.  I would prefer, therefore, to be able
>to apply the polnomial to any projection, rather than restrict it
>just to TAN.

I added a reference to Murray (1984) in section 4.1.2 yesterday to the
effect that ARC geometry better describes a Schmidt plate.  Of course we
were well aware of this, it's also mentioned in AIPS note 27.  However,
TAN with a third-order radial correction is just as good given that
third-order is required to account for differential refraction, etc.
Lasker et al. (1990) used a gnomonic projection for Schmidt plates for
the GSC (and so DSS) and refers to previous workers who did the same.
It appears to be the normal procedure in photographic astrometry.

Implementation requirements (iterative inversion, etc.) suggest that
there's only place for one polynomial in the WCS scheme, especially if
it's only needed in one place.

>	I contend that FITS WCS has to stop somewhere; to cover an
>infinite number of cases, it would have to be infinite.  So we should
>concentrate on the end-user; it needs to be accurate enough to correlate
>catalogs, images & maps, and do astrometry down to the requirements of
>target acquisition, fibre positioning, etc.  It serves the astrophysicist
>rather than the astrometrist.

Here, here!  And there's scope to add new projections or extend the
plate solution polynomial later if that should be required.

Also note that only the explanatory text has changed in this latest round.

Cheers, Mark