Some remarks on Rota's umbral calculus