The idea is good. It just doesn't work as efficiently as it could.
Using this method you could actually by pass the pivots altogether and just use the actual entities name.
But in what way is it inefficient. True, it generates lots of pivots, but it leaves the entities name free to be used for what it should be, doesn't seem to have a performance hit on any entities that use it, and the speed of FindChild is I am sure faster than parsing a long string as you have to parse the entire string in a linear fashion - the example creates 3000 properties for the cube and can retrieve any property almost instantly.
It also has the advantage of being able to allow array style properties in the new version i am working on.