After a brief historical introduction, we dwell on two recent experiments in the low-temperature, monoclinic phase of V2O3: K-edge resonant x-ray scattering and non-reciprocal linear dichroism, whose interpretations are in conflict, as they require incompatible magnetic space groups. Such a conflict is critically reviewed, in the light of the present literature, and new experimental tests are suggested, in order to determine unambiguously the magnetic group. We then focus on the correlated, non-local nature of the ground-state wave function, that is at the basis of some drawbacks of the LDA+U approach: we singled out the physical mechanism that makes LDA+U unreliable, and indicate the way out for a possible remedy. Finally we explain, by means of a symmetry argument related to the molecular wave function, why the magnetic moment lies in the glide plane, even in the absence of any local symmetry at vanadium sites.