Comments on "An analytic solution for an algebraic matrix Riccati equation"

In a recent paper by Shubert an analytic solution to a special class of Riccati quadratic matrix equations was proposed. The solution given there, based on completing the square, involves m = n(n-1)/2 parameters that are somewhat ambiguously defined. In this correspondence we point out that the ambiguity can be reduced by introducing orthogonal transformations. However, the form that the solution takes reveals that only a very restricted class of problems can be solved by the analytic method of the reference.