On Tue, 13 Feb 2001, Rachel Heery wrote:
> On Mon, 12 Feb 2001, Aaron Swartz wrote:
> >
> > > 6. Specification as to how to create DC instance metadata (including
> > > elements, qualifiers and schemes) in XML as per the schema in (4) and DTD
> > > in (5).(I do hope it would look the same!) This instance metadata to be
> > > XML compliant, but not RDF/XML compliant.
> >
> > It seems to me sort of silly to require us to do something, specifically not
> > in a certain way. What's wrong with RDF/XML?
>
> The reason we need XML (as well as RDF/XML) is that a lot of tools
> cannot handle RDF/XML. I agree there is nothing wrong with RDF/XML, that
> is not the issue. Altho DC supports RDF/XML can we mandate syntax?? I
> think not, therefore it seems to me DCMI should provide advice etc for
> XML expression.
I'm actually very much against any encoding of DC metadata which is not
possible to handle, at the choice of the services ingesting the metadata,
either as 'syntactic XML' or 'semantic XML' (i.e., RDF). The RDF
serialization which cannot be used either way is usually either poorly
coded or incorrect.
The 'semantic XML' or RDF arising from Eric Miller's musings March 13 last
year is very straight forward in that respect. Here in Lund, we have done
both. We have compiled such data to triples, and used various XML
software packages which doesn't have the slightest idea on what a triple
is.
Sigge
|