Comments on a paper of R. A. Brualdi: ``Notes on the Birkhoff algorithm for doubly stochastic matrices''