Some remarks on Landau's (macroscopic) phase transition theory