Such problems are inevitable whenever you try to represent a large number of symbols with a smaller number of symbols. It already happens in languages that natively use diacritics as a kludge to represent multiple different sounds with a single Latin letter, and is not helped by the fact that some of these languages lexically sort the accented letters with the base letter, some sort them after the base letter, and some sort them as a totally different letter, usually at the end of the alphabet. It's even worse when a letter pair is used instead of a single letter, and is similarly sorted as if it was a single unique letter. Fortunately, with the vast proliferation of language-agnostic computers, many such idiosyncrasies are slowly going away.