But my point is if you compute the order of the normalizer of an arbitrary Sylow p-subgroup, then you can count the number of Sylow p-subgroups (because the number of Sylows is the index of the normalizer of anyone of them in the group), which is (as far as I know) very difficult.