Realistic Satellite Rendering

Satellite brightness is highly variable, depending on many factors, some of which are not easy to determine:

1. degree to which object is in Earth's shadow
2. reflectivity / size of object
3. orientation of object (and thus which parts are illuminated by the Sun)
4. observer - object - sun angle
5. atmospheric absorption along observer/object path

Specifically item 3 is not obtainable for all but a few objects which have a known orientation, and the function of incident light angle to reflectivity is not well known for most objects.

Nontheless, satellite watchers maintain mean brightness data for 50% illumination, normalized for a distance of 1000km. One such source is Mike McCant's quicksat data (see https://www.prismnet.com/~mmccants/ -- It looks like this data is licensed appropriately to be redistributed with Stellarium - I have emailed Mike to verify this is OK, and await his response).

Implementation requires:

a) Earth shadow calculation. Ideally there should be some smooth fall-off while the object is in the zone where light passes from the Sun to the object through the atmosphere.

b) integration of brightness data into Satellites.json (with mechanism for updating this, although this might be a stand alone too for reading the json file and the brightness data and outputing an updated json file).

c) actual rendering as a point rather than an icon with the proper brightness.

Update: reply from Mike McCant regarding using his data:
----- BEGIN
No problem.

But you should be aware that there are two definitions
of "intrinsic magnitude" and mine is unusual definition.
The "Molczan definition" is pretty uniformly about 1.5
magnitudes fainter than the "McCants definition".
----- END

Note: the differences are elaborated upon here: http://www.satobs.org/brite.html

2010-12-10 Satellite eclipsing is now implemented in ~matthew-porpoisehead/stellarium/satellites branch.