This video demonstrates an algorithm that computes paths for a humanoid robot that should reach to an object on a table and that lean on the table to reach the object. The ultimate goal of this algorithm is to provide learning input for computing a memory of motion.