Rachel said:
> Absolutely! and this 'description' is what gives us some clues as to how
> we might 'trust' this schema, is it up to date etc. It plays the same role
> as an annotation might be expected to do perhaps? Tho there is no
> 'security' in that trust as yet, we are assuming that metadata tells the
> truth for time being :-)
Trust may be a part of the problem, but I don't think I'd got quite that
far...;-)
Leaving aside the appropriateness of using the "eor vocabulary", I think all
I was saying was that we may wish to make statements about instances of four
classes of object/resource:
(i) a namespace
(ii) a "vocabulary"
(iii) a document which is "at the namespace URI" (i.e. if it's a URL, the
page which that URL resolves to, which for DC will probably be a
human-readable document encoded as XHTML with embedded RDDL, if I understood
the consensus of a while back?)
(iv) an RDF Schema which describes the terms in a namespace in a
machine-readable form (which could be the document "at the namespace URI"
but won't be if we're putting an XHTML/RDDL page there).
I'm prepared to be persuaded that (i) and (ii) might be the same thing,
though I would like to be sure! See
http://www.jiscmail.ac.uk/cgi-bin/wa.exe?A2=ind0111&L=dc-registry&F=&S=&P=33
79
and especially Sean P's follow-up at
http://www.jiscmail.ac.uk/cgi-bin/wa.exe?A2=ind0111&L=dc-registry&F=&S=&P=37
93
But I think (iii) and (iv) are definitely distinct (from each other and from
(i)/(ii)).
So to make statements about instances of these four classes of resource (and
for others to know exactly what resource I'm talking about at any point), it
seems to me that I need separate identifiers for them which I can then use
as the value of an rdf:about attribute.
To use the namespace name/URI to identify all/any of these separate
resources rather interchangeably seems to me a recipe for horrible
confusion.
Pete
|