Since the node name is set after rospy.init_node() is called, there are many situation where we cannot know the node name (at import time for example...).
It seems useful to be able to get all PID, and later (when rospy.init_node is called), map these PID to names. Every function that needs a node_name can also take a PID instead.
So we rely on OS setting up PIDS, rather than ROS setting up names (although they are useful as alias)
Since the node name is set after rospy.init_node() is called, there are many situation where we cannot know the node name (at import time for example...).
It seems useful to be able to get all PID, and later (when rospy.init_node is called), map these PID to names. Every function that needs a node_name can also take a PID instead.
So we rely on OS setting up PIDS, rather than ROS setting up names (although they are useful as alias)
Not sure yet where that would belong...