So the planets are on rails? Like the old Renaissance mechanical solar system models?
It depends. You can define a planets orbit by state vectors in the cfg, in which case they are propagated according to realtime calculations of gravitational influences (that's the easiest way of simulating a planetary system on a computer programming wise, by the way). However, the whole thing gets highly unstable under high time acceleration.
This is why you can define a planets orbit by its orbital elements, in which case its position is propagated according to kepplers laws. Not entirely realistic (it's only a two-body simulation, other gravitational influences than the star and the planet in question are not considered), but absolutely stable under no matter how high a time acceleration, and even allow propagation over long periods of time without itteration.
Third, you can define the orbit by a custom dll, as is done for a lot of planets in the sol system. I.e. you can define the "rails" much more precisely, and you can take other influences into account and define the orbit accordingly (you can even simulate the errors in Mercurys orbit resulting from general relativity without actually simulating general relativity). Very precise during the course of probably several centuries, but it's still on rails.
3. Yes. You can easily "simulate" the planet's mass/size being changed. It is not practical but possible.
I guess you mean that you can simulate them having another mass by screwing with the gravitational influences on a vessel. Not very practical indeed. But if you ever find a way to actually change the physical size and mass or even add or erase a planet from a system, let me know. I would be highly interested in such a hack...
