On Wed, 2004-05-12 at 11:51, Pete Johnston wrote: > > Can the subject and object of a statement using the dc:identifier > property be the _same_ resource? I'd say it can, but it says something else than what was probably intended: <uri1> <dc:identifier> <uri1> means that the resource serves as an identifier for itself, which I suppose is trivially true. In other words: if I want to give an identifier for me, I can give myself (yes, the whole, physical me). But it's not a very interesting statement to make. What we're seeing here (in T1) is, I think, a nice example of a more general issue when expressing your own information in given frameworks, in this case RDF: You're using the framework as a *pure syntax*, ignoring the fact that there is built-in semantics in the model. This is different between XML and RDF, as XML has no (or almost no) built-in semantics, while RDF does. It's important that a mapping DC -> RDF or DC -> XYZ preserves the semantics of DC, while not conflicting with the semantics of RDF (or XYZ). This is not always so easy (LOM -> RDF is of course my best example of a problematic situation). I don't know if this discussion helped anyone, but it makes sense to me anyway ;-) /Mikael -- Plus ça change, plus c'est la même chose The more things change, the more they stay the same