The cell dynamical systems method of Oono and Puri [Phys. Rev. A 38, 434 (1988)] for modeling phase-ordering dynamics is reexamined. We show how rather puzzling features of this method can be analyzed on the basis of its relation to conventional partial differential equations.