The optimum value of the residual scaling factor c,
involves min(svd(A)) and norm(r), which are
usually too expensive to compute.
S = spaugment(A) without
a specified value of c, uses max(max(abs(A)))/1000.
In previous versions of MATLAB® product, the augmented matrix
was used by sparse linear equation solvers, \ and /, for nonsquare
problems. Now, MATLAB software performs a least squares solve
using the qr factorization
of A instead.