Comments on the high resolution photoelectron spectrum of H2O and D2O