Despite the fact that the method has become a standard topic in numerical  textbooks,  this  original paper is  still  widely read and cited. Apparently it remains the best reference on this algorithm. It's contents our so extensive that it contains the roots of the essential advances in this area up to nowadays: e.g. preconditioning to accelerate convergence of the algorithm, variations that can be used to solve systems of equations involving nonsymmetrical matrices, and evaluation of computational variants in order to choose the most numerically stable one.