the problem here is that text as the communication interface is not good for this. the model should be reasoning in the pose space (and generally in more geometric spaces), then interpolation and drawing is pretty easy. I think this will happen in some time.