**Abstract** : The presentation is focused on simulations, which ask questions about the validity of the numerical results. The guideline comes from a simple test case devoted to Maxwell's equations solver for which a homogeneous plasma is defined to fulfil the condition N=0 where N is the refractive index of this medium. The expected behaviour of the wave electric field obtained with a Laplace equation should be linear and exhibits always variations in time. To know why, some Dirac's mass with very low amplitude were added to see if we can recover similar shapes with a wave equation solver. Different spatial distributions of the Dirac's masses will be presented, finishing with a random distribution between −ε and ε giving again the linear behaviour as the standard solution. Now a new question appears what the impact of a white noise on the wave-electric field simulations? To test this, a resonance is a good candidate due the fact that a resonant condition is very sensitive to any perturbation. Thus a resonant condition was established and a white noise added to the refractive index. Analysing it, we see that the quality factor of the cavity was enhanced and increases non-linearly as function of the white noise level. To do that an extended precision is required for tuning the resonance at its maximum. After this excursion around N=0 we look at the impact of large variation of the refractive index. The study of such test cases was built based on the extraordinary mode in magnetized plasma. Using well-defined plasma density and magnetic field intensity profiles, the refractive index is able to reach high refractive index values. It is possible to reduce the group velocity by 2 orders of magnitude. As a consequence the computation time increases by 100. How to improve the computation time is open question. To finish PIC code simulations are shown, mentioning that Poisson's equation looks like the test cases with the Dirac's masses, and we conclude by addressing questions to the participants.