For the following, assume you're piloting a space craft. Assume a Euclidean coordinate system local to your space craft. Have that coordinate system oriented so that the center-of-mass of the ship is at the origin, the x-axis points directly out the nose of your craft, the y-axis points directly through the roof, the z-axis points directly out the right side of the craft, the w-axis points directly out
Typical video games where you're (mostly) locked on a 2-d surface (like car racing games, tank games, first-person-shooters, etc.) usually give you motion controls like, forward, reverse, turn left, turn right. (Some give you strafing, too... but not so much in the car and tank games.) So, you can move forward and back or rotate your coordinate system in the xz plane. With this one rotation, you can achieve any 2-d orientation.
Typical video games where you're (entirely) locked in a 3-d universe (like space flight games and flight sims and such) usually give you motion controls like: forward, roll left, roll right, pitch downward, pitch upward. So, you can move forward or rotate your coordinate system in either the xy plane or the yz plane. With these two rotations, you can achieve any 3-d orientation. Note: you'd still be able to achieve any orientation with xz and yz roations or with xy and xz.
So, we already see that moving from two- to three- dimensions, the controls from the previous dimensions are ignored. I suspect, however, that's a result of gravity being perpendicular to most 2-D games. I suspect that the fact that the 'y' axis is involved in both of the 3-D moves is because of the fact that gravity acts in that direction. And, space-flight games still play on the same idea that we have an intuitive feel for "down" (or at least for airplane controls).
So, if we were to move up to 4-d, what would be the most intuitive set of controls? I'm thinking that we'd keep the 3-d controls: forward, yz-rotations, and xy-rotations. Then, we'd add: yw-rotations. But, maybe xw-rotations would be more intuitive. I'm pretty sure that you'd be able to achieve any 4-D orientation with either set. But, what do y'all think... should we roll into the next dimension or pitch into it?
Another possible explanation for the 3-D flight controls is to think of an airplane with two wings, each with a flap. Rolling in the yz plane amounts to tipping the flaps on the wings in opposite directions (one up, one down). Pitching forward or backward amounts to tipping the flaps on the wings in the same direction (both up or both down). I can see two ways to try to extend this to four dimensions.
One is to say that the flaps can not only move up and down, but also ana and kata. Thus, you could do all kinds of things like: move one flap ana and the other down, move both flaps kata, move one flap up and the other flap up-ana, etc. This would probably result in controls very similar to tank driving games that give you two joysticks... one for the left-tread and one for the right-tread. In this, however, both joysticks would have a full 2-D range of motion. This would be easily implemented on modern joysticks which have two independent analog sticks. The only problem here is that to only have two flaps would mean that they're in the same plane. They'd be offset from each other in z, but not so much in w. Thus, it's somewhat like having flaps in xyz and only a rudder in xw, I think.
The other way to extend this is to assume, instead, that you've got one set of wings with flaps which are offset in +z and -z and can tip in a rotation parallel to the xy plane (these are normal airplane flaps) and another set of wings with flaps which are offset in +w and -w and can tip in a rotation parallel to the xy plane. Again, this would work decently with two analog joysticks. One would be the xyz airplane controls and the other would be the xyw airplane controls.
Thoughts? questions? comments? ideas? alternatives? cookies?