> Not a chance. Unicode and ISO 10646 collect in fact some unneeded crumbs > that way. Once a character is officially approved for inclusion it is > there forever, whatever we don't like it or not. Characters get > deprecated (i.e. usage discouraged) but never removed. This is the price > you have to pay for stability and usage by other specification (like > IDN).
Sorry, I did not mean removed from Unicode (I am aware of the rule there, which makes sense), I meant one character that is added to Unicode but that a given Stringprep profile would think desirable to remove. As for case folding, as soon as one adds a new case mapping in the Nameprep profile, one needs to upgrade clients at the same time as servers. Oops. Encoding the Nameprep version in queries would solve that, but obviously at the cost of some other meaningful info in the label, thus reducing the length of IDN names. YA
