Dear Pete, dear Andy, thanks for your patience. As Pete points out there is silence in the model about how one might access the related metadata in general - good. To provide a hint by means of "example" i don't see a problem with. Probably it's still ok with the abstract model, in case it happens, that (eventually) pointers to related metadata find their way back to the resource one originally started with. Hypothetical Example: A --dc:creator--> B Pointer starting at B ends in record C record C uses a dual DC property: "isCreatorOf" and contains: B --isCreatorOf--> A Please observe: I don't suggest to introduce dual DC. It's a hypothetical example for test purpose only. Best regards, rs