The encoding is UTF-16, but what it calls "characters" are code units http://unicode.org/glossary/#code_unit, not code points http://unicode.org/glossary/#code_point.