Science as heuristic search: Some comments on Simon's theory