Appendix F: Character String Comparisons
All metadata character string comparisons shall be done in conformance with the rules of the Unicode standard (http://www.unicode.org/versions/latest/) using the normalized form C defined in Unicode Technical Report 15 ( http://unicode.org/reports/tr15 ).
All XML ID comparisons and OID comparisons shall be compared in a case sensitive way.
Note: Latin alphabet case-sensitive NFC matching corresponds to byte string matching. The primary impact of this is for non-Latin alphabets. They need to be converted into normalized form before comparison. The TR 15 approach is consistent with the working documents of W3C, although W3C has not yet issued a balloted recommendation that Unicode normalized form C be used. See http://www.w3.org/TR/WD-charreq, http://www.w3.org/International/charlint/, and the current W3C draft (http://www.w3.org/TR/charmod-norm).