Nice. This helped straighten up my idea that svd is just eigenvalues for nonsquare matricies.