Rendering Glints: High res Nor mapped surface (Devs what do you think?)

So I watch this impressive video from Siggraph 2014 that touched on a topic that’s relevant to many of my projects.

Sorry if this has already been discussed, but I would like to know what Blender’s developers think of this:
Glints (or flecks, sparkles, what ever you like to call them) made by high resolution normal maps, coupled with a new sampling method

Here is the technical paper that goes with it.

I don’t understand the sample part (which is where the magic happens) but I do understand the high resolution normal map, as that was a method I attempted when trying to achieve a glitter paint effect. I knew my method was accurate, but I could not explain why my results didn’t turn out well.

The answer is in the paper, and I’m curious if something like this could/would be possible in Cycles (or BI) And what do the users think, would this benefit you or enhance your work?

Corona has this i think.

The importance sampling of glossy materials in Cycles has been improved recently (not in 2.71 though), have you tried using a recent build after that commit to see if it helps?