On Zariski geometries, structural approximation and rigorous Dirac calculus