Abstract: We present two ideas that help in the analysis of the dual affine scaling algorithm: sorting the slacks and taking a QR factorization of the constrains. Using these ideas, we prove that the iterates always converge. The proof holds if at each iteration we move an arbitrary fraction of the step to the boundary of the feasible region and it needs no hypothesis on degeneracy. However, it does not show that the iterates converge to an optimal solution. We present a new proof of convergence to the optimum for primal nondegenerate programs and for programs with two variables Ver menos