Better check for characters that aren't letters in ASCII, and assume that everything else is a letter (i.e. including Unicode). Not 100% correct, but simple and future proof.

And when it comes to the diaeresis, it's rare enough that it will create confusion for people who are used to German, Swedish, Finish, etc. (except for in words like naïve where there's no risk for confusion).

So will acute accents, which aren't incredibly rare. It's not unusual to see the accent preserved in fiancée.