> I certainly think it makes perfect sense, when one > reverse-engineers the > tree, but it might introduce an exception in place of > the one it > eliminates, if one insists on considering self > reference to be a cycle.
Sometimes you simply solve a mis-behaviour, though is certainly possible to make mistakes.
> The only funny part is that on level -1, where > there should be 2^-1 > nodes, or 1/2 of a node, there is one. But, it's > its own half parent, so...that kinda makes sense.
As for now, I guess this could be either related to the stress you are putting on "cycles", or to the fact that (maybe, I'm going to conjecture on this) zero is odd and one is even and so on, or both.
Indeed I can't see how in "our" framework the sub-root node in a tree leads to cycles...