Comment on time-variation of fundamental constants