One downside of this approach Ingo did not mention is that we’re now stuck with problems inherent to unicode such as han-unification and pre-composed vs de-composed normalization issues. But we can’t fix everything.
Unicode normalization was an issue for as long as Unicode existed, so we are stuck with it anyway. Pretty much like everybody else. And in many cases it is not even our problem anyway.
P.S.: I am still not convinced that Han unification is an issue at all.