The observations we made ( from NDR) regarding lots of optionality and flatness in e.g. the UBL:Address structure : The structure seems unusually flat ... We note that there's a lot of optionality of elements ... ... could be re stated in terms of good old fashioned database normalization basics . I wonder if it would be possible and valuable to restate some the definitions regarding N-th normal forms (from relational theory) in XML terms. Once they were restated the NDR could take a crack at a recommendation in those terms, e.g. we recommend 3NF (non-key columns depend upon the key, the whole key, and nothing but the key -- so help me Codd). This would make our recommendations more decidable. Regards, Bill