Just my .0000002 microcents.... Seems to me that poses fall into three categories in this connection. 1. Direct interaction with static physical objects, like playing a piano. In this case the pose would have to be done from the end, as in IK. Put the butt on the bench, put the feet on the pedals, put the hands on the keys, move the hands to the right keys. 2. Interaction with other parts of the body, as in putting on makeup or eating. In this case the two interacting parts would have to be brought to the same fixed point; pull the hand and the head to the meeting point, then retract both to original locations. 3. No interaction. Expressive gestures like a Gallic shrug or an Italian screw-you. In this case you could probably use the same joint angles universally, without worrying about the endpoint destinations. The first two types could probably be done with Python, using pre-formed motions of a guide prop, and setting the body parts to Point-At the guide props. See my "Hansel and Gretel" for a starting point. I think Bushi has made a universal pose-generator with a different method.

My python page
My ShareCG freebies