“Fighting for the acceptance of Bayesian networks in AI was a picnic compared with the fight I had to wage for causal diagrams [in the stormy waters of statistics].”