Into the detail of the "many worlds"

Anyone who reads about quantum physics would know that the rather improbable sounding "many worlds" interpretation of it has actually become quite popular amongst physicists.

But there's a review paper on arXiv from earlier this year that has a good explanation of the uncertainty and debate as to what the interpretation really means in terms of the "multiplicity" of universes.  Starting from Everett down, it's not actually obvious how it's meant to work.  

I don't think I had really appreciated the extent of the problem with the idea before.  The paper is not always easy to follow in every detail, but it is generally understandable and is well worth reading.  Here's the abstract:

Everett's interpretation of quantum mechanics was proposed to avoid problems inherent in the prevailing interpretational frame. It assumes that quantum mechanics can be applied to any system and that the state vector always evolves unitarily. It then claims that whenever an observable is measured, all possible results of the measurement exist. This notion of multiplicity has been understood in different ways by proponents of Everett's theory. In fact the spectrum of opinions on various ontological questions raised by Everett's approach is rather large, as we attempt to document in this critical review. We conclude that much remains to be done to clarify and specify Everett's approach.