I'm using spanish localization, and most of my documents are in spanish.

Spanish language uses very often accented characters: á, é, í, ó, ú, and also uses the dieresis (umlaut) on certain vowels: ä, ë, ï, ö, ü. Both situation may involve upper cases vocals.

Unfortunatly, spanish grammar says that when you write in upper case letters, you aren't obliged to put the accents on the vowels.

Besides that, it's very easy for a spanish person, and much more for a latin one, to forget accents and umlauts, so that when you are performing a search you never know if the words you entered aren't spelled different in some document.

Consecuently, I think that the only solution is to index accented characters words independently of the accents. I.e., you should index "organización" as "organizacion", and whether the user searches "organizacion" or "organización" or "organizaciòn" or "organïzación" search for the unaccented work.

I know that other languages have the same problem, at least italian, french, german.

Actually it seems that google uses this same strategy I suggest in this bug.