Question. I brought it up before... Why not find someone who will make a poseable "model" where you move the model and that generates pose information to the computer? Then you could use keyframing to generate movement. I know it's not quite the same theory but it would make doing initial poses so much easier!