Actually, a simpler solution, albeit perhaps not the best, would be to implement a form of weight mapping instead of using the split model/adjust joint zone method. With weight mapping, there's no need to break the model into seperate objects, which would cut down on the wierdness at extreme poses and couple that with the fact you can paint in how much an area is affected when moved then you'd have a joint zone killer on your hands. It's old technology I'd love to see in Poser some day but who knows what's in store for us in Poser 7, eh?