MR. KNIGHT'S letter inquires into the possibility of using computing machines to solve non-linear partial differential equations by the relaxation method. His suggestions imply two quite separate ...