Re: [office-metadata] Atom and document/feed IRIs

[Svante Schubert <Svante.Schubert@Sun.COM>]

Patrick Durusau wrote:
> I remember in the newsgroup days that all messages had unique IDs 
> (well, sort of, they were actually recycled after several years) which 
> is how they did threading.
>
> I am definitely leaning towards the IRI approach, mostly because it 
> makes the subject stable.
If we find in a RDF/XML an IRI, how do we know that it is describing an 
ODF element in our own package and not something else somewhere else in 
the world?
How is the connection to the ODF element in the document being made?

On first sight it seemed to me easier to put generated IRI on every ODF 
element, on the second I prefer again the xml:id approach, as one 
element in one certain document is being described.
If the document is renamed, moved, transformed to another data 
representation or copied, the metadata always describes the new entity 
of element in the ODF document, so the subject naturally changes as 
well. It rather improves consistency as the subject should represent an 
unique entity (odf element), otherwise owl:sameAs should be used.

Svante.