In solving diffraction problems, as far as I know, physicists treat the distribution of optical properties of the obstacle with wave equation, applying different simplifications (such as Fraunhofer limit).
I believe, that this treatment is compatible with the Huygens-Fresnel principle, which requests virtual secondary point sources to interfere in order to produce diffraction pattern.
My question is whether it is that different from the case, when these secondary sources are not virtual, but represent existing particles with nonzero contrast to the surroundings?
The latter is, as far as I understood, the basic requirement of the scattering theory (at least in Rayleigh scattering). Light scattering cross section is the (ensemble)averaged sum of phasors, whose "residence points" are small volumes of the actual particles with contrast. Since any scattering law can be separated (often, quite arbitrarily, but interconnected) in form factor and structure factor, why can't we assume, that there exist equivalent point sources, "structure factored" accordingly to particular scattering experiment, and that these are secondary waves' sources?
Is my train of thoughts legit?