Comments on "Minimal dimension realization and identifiability of input-output sequences"

In the above paper a method of realization from input-output sequences is considered. It is pointed out that the method can be used for realizing systems which are not strictly dynamic.
S=(A,B,C,D)
and the existing methods cannot be applied. It is further claimed that the proposed method is computationally more efficient for the state equation realization from a given transfer function matrix. In this correspondence we wish to show that these claims do not seem to be justified.