A small library that converts a number of reasonable compatible character encodings into Latin1 (or the subset ASCII or the superset windows-1252). Conversion may be strict or sloppy, i.e. letters may be 'unaccented' (eg. a with acute => a) or transliterated (e.g. double lower quotation mark => double quotation mark).
It's more convenient than iconv, recode or ICU if you're stuck with either ASCII, Latin1 or windows-1252. It's faster in unaccenting (for the supported encodings) then unacc, because it works on 8-bit characters. In case of licening problems it might help that library is LGPL'd.
More on sourceforge.net.