Comment on “Curvy-steps approach to constraint-free extended-Lagrangian ab initio molecular dynamics, using atom-centered basis functions: Convergence toward Born–Oppenheimer trajectories” [J. Chem. Phys. 121, 11542 (2004)]