On Mon, 12 May 2003, Roland Schwaenzl wrote:
> > On Thu, 8 May 2003, Roland Schwaenzl wrote:
> >
> > > Thinking about a suitable URI a little more i would suggest to use the
> > > DC-(X)HTML specifications URI. Probably that specification should eventually
> > > become an informational RFC, as it will not just talk about DC but also
> > > suggests coding for other metadata vocabularies in (X)HTML.
> >
> > Why? This recommendation is no more or less DC-centric than the XML
> > guidelines document is it? All the other DCMI encoding documents are DCMI
> > recommendations... why should the XHTML one be issued in a different way.
> > Either they should all be RFCs or they should all be DCMI recommendations.
> > Shouldn't they?
>
> 1. The former DC in HTML was issued as an RFC
Sure, but my guess is that at the time DCMI didn't have a process for
issuing recommendations itself?
> 2. rel="schema.something" is in fact a general suggestion, which allows for
> mixing of arbitrary vocabularies in HTML.
This might well be true. But documenting such a generic technique in an
RFC entitled "Encoding Qualified DC in HTML" doesn't sound very sensible
either? If you/we want to document a generic metadata technique then lets
issue an RFC entitled "Using the XHTML link tag to indicate metadata
namespaces"?
> Language: In the "abstract model" values of properties are required as
> "literal string", but URIs in links are adressed as URIs for "another resource".
(Again, it would be helpful, to me, if you could make concrete suggestions
for alternative text here!).
To be honest, I don't see a major problem here... I'd be interested in
other people's views. Should the abstract model use 'string' rather than
'literal string'?
> "link" Attributes: Is a dc-application assumed to ignore all other "link" tag attributes.
No. I'd turn this statement on its head. A dc-application must handle
the rel, href and hreflang attributes and should handle all the other
attributes. I.e. a dc-application is assumed to support everything
specified in the recommendation but may also support other aspects of
HTML/XHTML as appropriate.
> is "rel=copyright" the only built in "link" type, which is to be honoured by a dc-application?
It appears, to me, to be the only builtin link type where there is a
semantic overlap with DC. However, I'll re-word based on Chris's
suggestion.
Andy
--
Distributed Systems, UKOLN, University of Bath, Bath, BA2 7AY, UK
http://www.ukoln.ac.uk/ukoln/staff/a.powell +44 1225 383933
Resource Discovery Network http://www.rdn.ac.uk/
|