As Aaron and many others pointed out: One URI can serve several purposes. One is: They imply other URI's xmlns:dc="http://purl.org/dc/elements/1.1/" automatically from the specification gives "http://purl.org/dc/elements/1.1/title" when we use dc:title in an RDF/XML record. That's how it is. The current resolving (redirect) of "http://purl.org/dc/elements/1.1/" is to "http://dublincore.org/2001/08/14/dces#" . I don't think "/ #" substitution is dictated by the definition of "purl" - or?. The consequence of the seemingly innocent change "/" to a "#" at the end, is that we retrieve always the same object: Regardless whether we call the namespace URI or "http://purl.org/dc/elements/1.1/title" - In case i'm right about purl resolving, there is a plethora of things we could discuss under the header "to what resolves..." Anyhow: An RDF Parser should get from the the result of retrieval of "http://purl.org/dc/elements/1.1/" at minimum what it get's right now about dc15. [Analogously for terms and types and ....]. ------------- RDDL: My understanding of http://www.openhealth.org/RDDL/rddl.rdfs# [1] is, that RDDL has a preferred interpretation in RDF. This is impression is supported further by http://www.openhealth.org/RDDL/rddl2rdf.xsl [2] This causes a problem thanks to substantial semantic overlap of RDDL with DublinCore - recognizable for a human. On the other side form the viewpoint of an RDF application taking [2] into account, RDDL currently appears as a metadata language unrelated to DublinCore. [1] and [2] are declared by http://www.openhealth.org as non-normative examples. I have no idea what a non-normative example possibly could be. It would be desirable for a metadata registry to provide machine understandable relations - I'm not sure about whether the DC-registry people have considered this kind of issue. I think it's a fairly real one - as this kind of behaviour RDDL shares with some other metadata languages with preferred mapping to RDF. --------- Cheers rs