TODO

1.	Have to get photometric correction working.
2.	Have to eventually change default ostream for FITS routines.
3.	Implement GUI.
4.	Test the remaining projections and implement grid routines for them.
5.	Implement signal handling.
6.	Add despiking routine.
8.	WORK ON DOCUMENTATION.
9.	Create HST-FITS file handling routines.
11.	Add PostScript support.
12.	Add JPEG and additional image format support.
13.	Add support for other types of FITS formats. (e.g. Ascii Header Ext.)
15.	The input image input file entry should be slightly reorganized.
16.	Implement map to map conversion routine(s).
22.	Correct Perspective projection constructor to use new transformation
	matrix procedure.  Correct Perspective projection grid creation routine
	to use new perspective routines.
24.	Create ".el" lisp module for Emacs/XEmacs MaRC input file mode.

-------- Following from Joe Harrington -------
27.
Minnaert and IR Minnaert (no incident contribution) -- see Helfenstein
automatic limb centering -- see jh
automatic registration -- see jh
options for composites that don't use averaging:
        use best distance from terminator or limb in source image
        use best mu*mu0
input  image masks - 0 or 1 depending on validity of pixel
output image masks - count of contributing input images per pixel

Output options:
options for different data types for each created object
option to write any item as: primary array image, extension, and/or sep. file
grid specifiable like MU/MU0 (spacing in lon and lat, origin location)
MU/MU0/PHASE/GRID bscale and bzero keywords, e.g., to fit in 0-255
Input file:
averaging method defined per-plane as override to global def
entire input file to appear in output as comment
output file should identify which item in the input file it is
simple_cyl -> equidistant_cyl or equi_cyl
sample -> naxis1
line -> naxis2

PROBLEMS:
ragged edges on maps?
HST extension misunderstood
equinox keyword

Make all the docs install into ${prefix}/doc/MaRC-version/, which
seems to be the emerging standard.   (......yeah?)

Perhaps mumu0.test.inp could be used in a make check?  You could
define a very small output image and include it in the package, and
cmp it with the one that gets made by make check.

Docs for the library?  Yeah, I know, pipe dream... :-)

read any planetary parameters from FITS keywords if not in map input file

---------------------- End jh suggestions -------------------

27.	Implement map auto-centering (i.e., automatically determine what range
	of latitudes and longitudes to map.
28.	Move Centric{Radius, Latitude}, M, N, etc. functions to BodyData class.
	These functions should be declared virtual.  An "OblateSpheroid" class
	will then implement these functions.  This change will be used to allow
	additional body models to be used, such as spheres and "spuds."  The
	"spud" model would be something similar to the one used in the
	spudproj program which is now being used to map asteroids.  All map
	projections should then have different versions of their prospective
	map equations for each type of body, namely: OblateSpheroid, Sphere and
	a general body type (such as spuds).  An attempt should be made to
	implement a general ellipsoid class (is this worth it?).
30.	Add support for SPICE library.
31.	Move many of the "valid value" tests to "appropriate" locations.  For
	example, tests for proper standard latitudes should be moved to
	projections that use standard latitudes, perhaps even to the
	underlying base projection classes, such as Conformal and Equivalent.
32.	Multithread the MaRC library and the MaRC main program.
33.	Add command line options via the popt libary.
34.     Fix all the FIXMEs in the code.