id like to see a planetary capable flight model, with a basic aerodynamics engine, and a proper implementation of gravity. but then again im not sure if freespace can handle missions of that scale. definately ship tweaks are needed. it would be grossley unrealistic to be able to fly out of a planet's atmosphere at 80 meters a second.
Implementation of gravity:
Put an acceleration vector on every ship that makes them acceleration into one direction (DOWN). Easiest if the gravity field can be seen as homogenous, more difficult if the gravity field is spherical. On FreeSpace scale, we can just construct a skybox and then make the ships fall.
I doubt that would be very difficult.
Implementation of aerodynamics:
This is the more tricky.
We need situation-variable acceleration vectors that are defined by
air speed (v), angle of attack (AOA), air density (constant, mission specific? derives from pressure and temperature) and series of ship-specific constants (namely drag coefficient and lift coefficient.
Also, we need proper angular momentum.
Now we have the means of generating the results of four forces that affect an airplane:
Weight - shown to us as gravitational constant
Lift - counters the weight; is a result of air speed, angle of attack and lift coefficient
Thrust - already ready; seen as forward acceleration
drag - seen as backwards acceleration. Always on 180 degree angle from the air speed vector. Is a result of air speed, angle of attack and drag coefficient.
This is the most easily implementable aero flight model. If you want to get anal with it, you can of course implement additional force vectors that affect certain parts of the ship (ie. control surfaces) and thus give the ship angular acceleration and stuff like that.

Actually, as long as we aren't craving for ultimate realism, it would probably be rather easy to implement these modest things into the engine. We don't even need to start counting the lift and drag coefficients or anything, just meddle with them and assign appropriate values for each ship. Though I'm not a coder, so I can't really estimate the amount of work needed to do with this. It would sure add multiple flight parametres and variables and constants, and probably add awful lot of conditions to prevent div by zeros and whatnot.
For a very, very simple model, that is needed. Any coder want to try this? After all, it's quite basic vector physics in itself.
