[fitswcs] Polynomial projection

[William Thompson]

While playing around with the GRISM projection, the thought occured to me, why 
not have a polynomial projection?  I know that Paper IV discusses polynomial 
distortions, but projections are simpler than the distortions implementation of 
Paper IV, and generally preferable when possible.

The new projection would be -PLY, and would be defined such that

	S = S_r + w + A_2*w^2 + A_3*w^3 + ... + A_20*w^20

The various A_k values would be stored in the PVi_ka keywords.  By default, 
PVi_0a=0 and PVi_1a=1, and no other values would be allowed for those two 
keywords.  The default PVi_ka with k>1 would be 0.

The use of -PLY would be like that of the -LOG projection, e.g. CTYPEia = 
'WAVE-PLY'.

It's probably too late to include this in Paper III, but it might be included in 
Paper IV.

-- 
William Thompson
NASA Goddard Space Flight Center
Code 612.1
Greenbelt, MD  20771
USA

301-286-2040
William.T.Thompson.1 at gsfc.nasa.gov