[fitswcs] Grisms

[William Thompson]

There are 7 parameters which make up the Grism projection, i.e.

	G		Grating ruling density
	m		Grating order
	alpha		Angle of incidence
	nr		Index of refraction
	nr_prime	Derivative of nr with wavelength
	epsilon		Angle between normal and dispersion plane
	theta		Angle between reference ray and camera axis

When I go through the math, though, it seems to me that these collapse down to 
only four independent quantities:

	G*m/cos(epsilon)
	nr*sin(alpha)
	nr_prime*sin(alpha)
	theta

Unless I've made a mistake anywhere, this should be discussed in the text.  If 
one tries to establish the 7 parameters by means of a least squares fit, one 
ends up with a degenerate problem.  I anticipate that fitting would be a common 
procedure--that's certainly what we do today, except that we usually fit 
polynomials.

I have no objection to keeping the original set of parameters, because they're 
strongly connected to physical properties of the instrument, and relatively 
simple to understand.
	
-- 
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