Comments on the fractional step method