Rationalize ACLs vs private column in some tables. (pro_identifier_types, prs_biodem, prs_addresses)
History models. (changelog, _archive tables, deleted columns)
Should calculated tables other than prc_persons all be part of BR?
Should BR and PR be combined?
Will this data model scale for large quantities of data?
Should SORs be able to provide name data as name,type={preferred|maiden|typo} to help OR elect best name?
Is multiple stage deprovisioning supported properly in the data model? ie: you're expired but not deleted, oh, now you're deleted (this might just be two business rules)
Implementation
Should interfaces support "-n" option to see what happens before a change is submitted?