That's an interesting trick for dealing with uniqueness across case-sensitive and case-insensitive cases. I wonder if they were having issues disambiguating IDs at some point and instituted this, of if they do it to interoperate with legacy systems lacking support for case-sensitivity.