Left: sketch of the real geometry of the HII region in the presence of accretion (based on models by Peters et al. 2010a,b). Right: our approximation by a spherical HII region filled with electrons and ions, but no dust. The PDR and the molecular gas surrounding the HII region are not represented in the sketches. Since the accretion disk does not cover a large fraction of the volume, and does not shield a large part of the spherical surface, we deem the approximation by a spherical region a reasonable approach, which greatly facilitates the calculation, without deviating too much from the real situation. While Peters et al. (2010a,b) show that the situation is dynamic and the shape varies on short timescales, following the true dynamics is beyond the current computational capabilities. We believe that, while the current static model is notstrictly correct, it still provides valuable insights into the process, even though it eventually will have to be superseded with a model incorporating the dynamics.