rdf canonicalization question

Hi,
I briefly skimmed through the stuff in the archives about C14N canonicalization of RDF. But I don't see anything about a solution to the problems which graph isomorphism stuff seems to present.
Canonicalisation for RDF presumably is supposed to collapse semantically equivalent graphs into one textually identical version. Has anyone thought up a solution for what to do about canonicalizing semantically equivalent graphs which are expressed say with different node ordering, or with different configurations of bNodes? I know there is a good paper on determining the identity of graphs, but has anyone done any work on making identical graphs textually identical?

Thanks

Giles

_____________________________________________
Giles Hogben
CyberSecurity Unit
Institute for the Protection and Security of the Citizen (IPSC)
European Commission - Euratom Centro Comune di Ricerca
Ispra,   Italy