Emulsion films (developed by F.Powell) for example show that elementary particles follow a straight trajectory staying in a film. The reason was clarified by Francis Mott in his famous 1929 paper about alpha tracks. It implies that at each weak interaction the wave collapses (decoherence) to a packet moving in a straight line. I don't understand why this can not be the same with a photon in thin film of weakly absorbing substrate on a plate (by example the film only rotating the polarization plane of the light). As far as I have been told this is a case of forward coherent scattering, which is some sort of interaction which doesn’t change the impulse (or at least its direction). So the first condition of a weekly interacting particle is met. The second is also met: the photons must be scattered exactly forward. Now I think that a photon can not follow a straight trajectory in a thin transparent film even provided the source has emitted the photon inside the film with impulse parallel to the film surface. (lambda greater than thickness). But according to the alpha tracks and the Mott analysis this must be so. Why?