Comment on "efficient time-domain beam-propagation method for modeling integrated optical devices"