Discussion of ``One-step sparse estimates in nonconcave penalized likelihood models'' [arXiv:0808.1012]