Yes, that's from Max Tarpini, another Italian, the author of the well known ctrl shader collection, who also helped me for my shaders.
In theory they also should work in Soft 2011, since you can import the shader declaration by their mi files.
I discovered them after I was almost done with mine, but I kept on going, since we have a different approach to the problem, in particular in the way of generating the points. I think he addresses better than me some particular problems (like sss) where I leave more work (and maybe choice) to the user. Anyway, good to have more than one option.
Both methods were made possible by the new (from mr 3.7) map api, a way of storing and reading very large arrays of points, more efficient and fast than the old kd-trees. Mental is also migrating their point set formats (fg maps, photon maps, etc) to a unified map format, which is a very promising ideas. Beside being faster, we shader writers will finally have access to those data set. So, in theory, with the next versions it will be possible for instance to read/edit/rewrite fg/photon maps, which I think is great. A photon map may be physically correct, but so many time you don't care about that, you would just like to be able to "paint" a map and let the surface shader read it.
I am open to suggestions (and so I think Max is on the mental forum). For instance, I am thinking about using the input pdc only to define the points position, but to recompute the mrmap color and normal at each frame. This way, the user could stick a given points distribution over the surface, so fixing the current main limitation I have now (distribution and flickering).
Ciao