Comments on Tisdell's critique of Nerlove-Waugh theorem concerning optimal advertising