A correction for the bias of maximum-likelihood estimators of Gumbel parameters — Comment