How does the concept of an ideal thin lens help to simplify calculations and make predictions about the behavior of real lenses?