Summary
While we weren't able to accomplish exactly what we had set out to (the USB gravity joystick), we did end up with a project that was close relative of the original plan and interesting in and of itself. A mouse over TCP/IP is fairly unique (VNC is the only other instance we can think of) and to have it accelerometer based as well adds to its mystique.
In the end, we feel that some of the poor decisions that we made in our initial specifications were actually invaluable learning aids. We were forced to figure out exactly what was feasible, and what was simply going to be too hard to implement. We also had to determine a new tact that would help us avoid the newly identified pitfalls while still moving us towards our final goal. All in all, we are all proud of what we accomplished (even if we are feeling a bit frazzled right now).