Two issues with this latest draft:
Issue 1 -- I'm concerned that proposed method for linking to related
metadata my encourage ambiguities. For instance, the gray-box example in
"Section 7 - Linking between metadata records" includes the following:
<dc:creator xlink:type="simple" xlink:href="#andy">
Andy Powell
</dc:creator>
...
<person id="andy">
<vcard:fn>
Andy Powell
</vcard:fn>
<vcard:org>
UKOLN, University of Bath
</vcard:org>
<vcard:email>
[log in to unmask]
</vcard:email>
</person>
The relation between the linked metadata (i.e., the vcard information) and
the metadata contained by the dc element is unclear. In this case there's
redundant information, generally not a good thing. Worse yet, consider if
the string content of the dc:creator element in the above example where
"Andrew Powell" (no offense, Andy). It would then be unclear whether Andrew
Powell or Andy Powell were the preferred version of the name. Does the
vcard metadata takes precedence, or is it really intended that both forms of
the name are to be used? What's the relationship between the two forms?
Wouldn't it be better when linking to related metadata to require an empty
element (i.e., more like, dare I say, the way RDF does it). Thus:
<dc:creator xlink:type="simple" xlink:href="#andy" />
...
<person id="andy">
<vcard:fn>
Andy Powell
</vcard:fn>
<vcard:org>
UKOLN, University of Bath
</vcard:org>
<vcard:email>
[log in to unmask]
</vcard:email>
</person>
Issue 2 -- I'm concerned that we're struggling here to reinvent the wheel
just to avoid the dreaded "rdf" namespace prefix. Both Aaron (earlier note)
and Chris (below) have pointed out that this document is close in many
respects to the "Expressing Qualified Dublin Core in RDF / XML" guideline,
mostly with some of the names changed to avoid mention of RDF.
Intellectually you still have the equivalent of rdf properties (DC elements)
and rdf subproperties (DCQ refinements). The linking of metadata could just
as easily be done using rdf:Resource and rdf:ID attributes as XLink style
attributes. The major difference is the substitution of dcxml:scheme
attribute for the rdf concept of typed nodes when handling DCQ encodings.
While I don't have a problem with this alternative syntax, it seems a shame
to establish a competing way of expressing DCQ in XML just for this one
small and arguably trivial difference. If you account for that one
syntactical difference all of the examples in the document can be rewritten
with just semantic differences and run through the SiRPAC or the W3C RDF
Validator.
I'd much rather see this rewritten as a specific application of the
Expressing Qualified Dublin Core in RDF/XML guidelines, analogous to the
"application profiles" that have been written to help implementers of DC in
particular communities. The range of syntax options available in full-blown
RDF is intimidating. Rewritten consistent with RDF, the approach of these
"Guidelines for Implementing Dublin Core in XML" would essentially prescribe
RDF syntax choices and make the process of expressing DC/DCQ in RDF/XML much
easier. This should be sufficient for most implementers. If there were a
larger divergence in underlying concepts, that would be another matter --
but given how close this new document already is to being intellectually
consistent with the RDF/XML guidelines (albeit more prescriptive), it seems
disingenuous to present it as a fully independent and separate approach to
expressing DC/DCQ.
Tim Cole
University of Illinois at Urbana-Champaign
----- Original Message -----
From: "Chris Croome" <[log in to unmask]>
To: <[log in to unmask]>
Sent: Sunday, February 03, 2002 11:51 AM
Subject: Re: Guidelines for implementing Dublin Core in XML
> Hi
>
> On Sun 03-Feb-2002 at 09:17:36 -0600, Aaron Swartz wrote:
> >
> > I think all you'd really need to do is change <metadata> to
> > <rdf:Description> and use the qualifier style specified in the
> > DC-in-RDF spec.
>
> Good point, if you take the first example in the document [1] and do a
> RDF version (I think the attached file is more or less right), there is
> not really that much difference.
>
> Following is a result of a diff between the two versions (lines starting
> with < are from the XML version and lines starting with > are from the
> RDF version):
>
> 3,7c3,5
> < <metadata
> < xmlns="http://myorg.org/myapp/"
> < xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance"
> < xsi:schemaLocation="http://myorg.org/myapp/
http://myorg.org/myapp/schema.xsd"
> < xmlns:dc="http://purl.org/dc/elements/1.1/">
> ---
> > <rdf:RDF xmlns:rdf="http://www.w3.org/1999/02/22-rdf-syntax-ns#"
> > xmlns:rdfs="http://www.w3.org/2000/01/rdf-schema#"
> > xmlns:dc="http://purl.org/dc/elements/1.1/">
> 8a7,8
> > <rdf:Description rdf:about="http://www.ukoln.ac.uk/">
> >
> 24,28c24
> < <dc:identifier>
> < http://www.ukoln.ac.uk/
> < </dc:identifier>
> <
> < </metadata>
> ---
> > <dc:identifier rdf:resource="http://www.ukoln.ac.uk/" />
> 29a26
> > </rdf:Description>
>
>
> Chris
>
>
> [1] http://www.ukoln.ac.uk/metadata/dcmi/dc-xml-guidelines/
>
> --
> Chris Croome <[log in to unmask]>
> web design http://www.webarchitects.co.uk/
> web content management http://mkdoc.com/
> everything else http://chris.croome.net/
>
|