Domenico,
The mirror problem resolves with c in the medium frame both on approach and after reflection, because the 'transition zone' must of course work both ways. It is equivalent to a 'fluid dynamic coupling' of inertial frames. At the boundary there are electrons at rest in BOTH media frames, with turbulence between, but all re-emitting at their own c. This applies to ALL media. Only the extinction distance changes (with 'birefringence' apparent during the change).
I think this has proved too difficult dynamic visualisation of most readers, so introducing theoretical speculative neutrino's would just confuse more. We need to assimilate the concept that there is BOTH real and APPARENT speed, so apparent superluminal speed is allowed with no violation of c.
At the (twin-slit) back screen, there is no co-motion, so the (probably part re-blended) arriving 'photon' or wave energy charges the particles, and as each reaches threshold energy it re-emits, forming the small 'dot' pattern. The intensity of combined interacting waves from the 'slit edge' emissions changes across the screen according to Huygens construction to form the interference pattern. This can simply be shown in simple experiments I've done myself. By moving the back screen position during the experiment. A 3D interference picture can then be built up in the space behind the slits. You may have read this last year; http://vixra.org/abs/1007.0022
Does this accord with your thoughts? What it does seem to do is unify all physics and derive curved space time from QM by changing the optical axis of re-emissions ('refraction'). Is that emerging for you, or can you see any shortcomings?
Also the little maths I give has been challenged. Do you have a view on that?
Many thanks, and Best wishes
Peter