Modifying the variational principle in the action integral functional derivation of time-dependent density functional theory

According to a recent paper by G. Vignale [Phys. Rev. A 77, 062511 (2008), arXiv:0803.2727], the problems arising in the original derivation of time-dependent density functional theory (TDDFT) based on the Runge-Gross (RG) action-integral functional (AIF) are due to an incorrect variational principle (VP). This argument and the proposed modification of the VP are critically analyzed. The more fundamental problem, though, is the indefiniteness of the RG AIF. In contrast to a widely held belief, that indefiniteness is not eliminated in the variational procedure, which unwittingly is corroborated by Vignale's initial point.