Computing QED path integrals if fraught with issues such as “Is the detector direction sensitive?”. The classic two slit experiment reports photon strike densities on 2D screen which could be replaced by an array of directional photon detectors. An omni-directional detector replacing the whole screen at some point in the plane of the screen, will see a different intensity than the detecttor it displaced. It is the same reason that polar regions, even in their summer-time, get less light per square mile that equatorial regions. Ray tracing, as a graphics discipline, must get these issues right and so must QED—thus this digression.
This program is a MonteCarlo calculation of a light source at (0, 0, 0) shining on a flat scattering barrier at (x, y, 1), one unit away. The barrier transmits all light but scatters it a la Lambert. For X > 1, shine(X, z) returns the illumination falling on another parallel plane at (x, y, z) where x2 + y2 = X2.