At the heart of the computation of model atmospheres there is the so-called Stellar Atmosphere Problem, which consists of the self-consistent solution of the radiative transfer equations under specific constraints. The amazing progresses achieved in the field since the 1970s are due to both the dramatic increase of the computational facilities and the development of effective numerical algorithms. The purpose of this review is to draw attention to some methods, alternative to those that are mostly used nowadays such as the ALI methods. The improvement of the latter has been brought about by mathematical refinement, whereas the former are the result of a careful analysis of the physics of the problem. Rather than attempting an exhaustive presentation of these novel methods, which would be out of place here, the prime aim of this article is to sketch the main guidelines and to stress that it is always the physics itself that dictated the most effective algorithm.