We reexamine the classical principal-agent problems and place them in a new perspective. Our motivation comes from two observations in the real world that fail to be captured by the classical principal-agent models: the agent's action space is rich and optimal incentive contracts tend to be...