Note from conversation with John S. 1530h, 2 Oct 96:

- mistrusts results for star density near 0 RA, 0 dec - possibly a
wraparound problem (what is -1 MOD 5? 4 or -1?) - suggested taking 
(-1 + 5) MOD 5;
- recommended Graphics Gems for quick algorithm for finding smallest 
circle enclosing 3 stars;
- stars closer than 3 deg can interfere depending on orientation of
joining segment wrt grating rulings;
- if O3 cuts off above 40 km and apogee is at 80 km, can see stars down
to 96 deg ZA without attenuation from apogee; this expands the number of
stars we may be able to use for tomography;
- if spectra are dispersed in direction of payload rotation then _any_
rotation during exposure blurs spectrum beyond recovery - have to
disperse spectrum perpendicular to expected motion and try to recover
afterward by integrating along motion - having 0th order in the image
would make this easier;