Recently, I am working on a set of realisticly-colored maps of Jupiter. One issue of Jupiter maps is to get the latitude-structure of the cloud patterns right. For an oblate planet like Jupiter, this is especially difficult. As far as I know, Celestia just squeezes Jupiter into an oblate spheriod, without considering the consequences on the 3D body. For an extremely squeeced object, the terminator is clearly wrong. This is less obvious at Jupiter.
Also the coordinate system just gets sqeezed. One could overcome this problem by using a correctly projected map.This Cassini map of Jupiter
is projected planetocentric. When using it in Celestia, I get a discrepancy when comparing it to actual Cassini images. So this projection is not suited for Celestia.
In contrast, this map by Björn Jónsson
is projected planetographicly. This projection is consistent when comparing to Cassini images.
Is there an easy way to reproject textures from planetocentric to planetographic? The better solution would be if Celestia used an accurate 3D-model of oblate (or even triaxial) objects, so the problems mentioned above wouldn't exist. We could use planetocentric maps then also. Or how about a parameter determining the coordinate system of the planet and thus the map projection.