[ODE] point-on-plane constraint

[Jon Watte]

> I don't know how this could work, because the standard ball constraint
> already removes the 3 DOF for the xyz movement, but how could I
> archive the required motion then?

You only want to remove one degree of freedom (along the axis of the 
plane normal). Thus, you want to take the ball joint, and remove two 
of the constraints that it currently imposes, rather than adding more 
constraints to it. Never having written an ODE joint, I don't know 
what the details would be -- it might be that the movement necessary 
in a "constrain to plane" joint is actually different enough that 
it's a whole new joint type.

Cheers,

			/ h+