Curt Arnold Parsing a non-Unicode normalized document should result in Unicode-normalized content if normalize-characters is true.. 2004-03-31