The BACON algorithm, introduced by Pat Langley and Herbert Simon etc. were meant to automate scientific discovery -- producing causal explanation to variations in given data.
It was found, in particular, to have been able to "discover" the laws of planetary motion, but there were criticism concerning whether or not its achievement could be consider true scientific discovery.
The one I remember was that--the data given to the algorithm are pre-processed by human operators so as to include mostly relevant factors of the world. For example, the distances of planets from the sun was given, showing that the human operators implicitly understood the importance of this variable for the laws of planetary motion.
I was wondering if there were other significant criticism of the algorithm--either in general as a automaton of discovery or concerning its particular feat with planetary motion.