Revision history [back]

rospy.spin() will effectively go into an infinite loop until it receives a shutdown signal (e.g. ctrl-c). During that loop it will process any events that occur, such as data received on a topic or a timer triggering, when they occur but otherwise it will sleep. You should use spin() when your node doesn't do anything outside of its callbacks.

rospy.sleep() will do the same as rospy.spin() but only for the length of time you specify. It is used when your node has to do some kind of regular processing in addition to responding to topic events, timer events, etc. For example, a node controlling a motor will need to do some processing at regular intervals to set the motor's desired speed.

(Technically, what both are doing is sleeping your main thread and allowing other, internal threads to do their work.)