“…The computer models, in fact, seem superior, in important respects, to our traditional models. First of all, they have been specified precisely-this is not at all trivial; recent attempts to put such theories as Hebb's and Hull's into precise, testable form have met with major difficulties (Dunham, 1957, Rochester, Holland, Haibt, & Duda, 1956. Second, they have actually been examined for their consequences (through the running of the program) and tested as to their success.…”