Allowing a pin-slot joint to follow a spline [and not just a linear straight line] is essential as currently there is no way to properly represent designs that use curved sliders which are A VERY COMMON MECHANISM [without resorting to turning on slow calculation intensive collisions]. Most of my designs have more than one of these and even the most hardcore workstations would struggle to handle these designs due to this issue.
An implementation of this type would have to discern the difference from a 2D and 3D spline in that a 2D spline could use a revolution based axial "pin" type of pivot, but a 3D spline would likely require a point 'ball' type of pivot since a coplanar reference would not exist to define the 'pins' axis [commonly perpendicular to the plain] and defining an arbitrary plane reference does not make much sense from a mechanical standpoint [unless defining a planar reference at each point in the spline]. Because I have not used any 3D slots in my designs and as a result have not heavily considered the full implications of such a thing, I cannot be certain if it would be better to not allow 3D splines or to allow them and change the pivot type to 'ball'. My immediate notion would be to allow them so it isn't an issue down the road for those that would use it and since the implementation of this would just be one less restriction on the pivot axis.
In the 2D case it might also be handy to allow a point 'ball' type of pivot [instead of only an axial pivot] which, for certain designs, would reduce the number of additional joints to be defined.
