For large-scale sparse saddle point problems, Peng and Li  have recently proposed a new alternating-direction iterative method for solving nonsingular saddle point problems, which is
more competitive (in terms of iteration steps and CPU time) than some classical iterative methods
such as Uzawa-type and HSS (Hermitian skew splitting) methods. In this paper, we further study
this method when it is applied to the solution of singular saddle point problems and prove that it
is semi-convergent under suitable conditions.