On Fri, 20 Jul 2001, Rebecca S. Guenther wrote:
> One is "role". We are talking about it as an element refinement. They use
> an element attribute "role= " and the MARC relator codes (not terms, but
> codes).
Yes, the modelling is different here. I'd expect
<dct:Photographer>...</dct:Photographer>
where they are proposing
<dc:Creator role="pht">...</dc:Creator>
:-(
> The other is "file-as". Publishers have a requirement to display the names
> as they appeared on the piece, and they are saying to use that form for
> Creator and Contributor. However, they also want the normalized form
> expressed for indexing purposes, so they have an attribute "file-as". The
> DC Guide tells people that best practice is to use the inverted form. But
> what if you need both? And they do want to identify that one is the
> display form and one the normalized form.
My personal view is that what they've done is fine. It's an application
specific solution that nobody else will understand - but the
implementation looks reasonably sensible to me.
> So, how to advise them? I'm gathering opinions. However, I have to say
> that they are very reluctant to change anything in their spec because of
> backwards compatibility. They may not agree with anything I recommend. And
> they also need an answer right away and can't wait for a long approval
> process. So my inclination is to leave it for now, but wanted to get
> opinions.
Nothing they've done looks drastically bad - particularly given the
minimal amount of information we've given them to work with - see below...
so I tend to agree with you.
A couple of other comments...
They use upper-case first letters - I assume that these days we might
recommend lower-case. I.e. we'd recommend <dc:title> over <dc:Title> ??
But the fact that I have to phrase this as a question rather than a
statement indicates that I'm not sure that the DCMI knows the answer to
this so, not suprising that people outside the DCMI have to make stuff up!
:-(
Under dc:Format they say... 'An enumerated list of formats for this
content is being developed by the Dublin Core.'. This isn't the case.
Either they are confusing it with dc:Type ?? or they are confusing it with
MIME types ??
> FYI: they use an XML DTD. And I guess I have no guidance for them on
> expressing DCQ in XML (or am I missing something-- I didn't see any
> proposed recommendation for that).
I disagree with Stu on this. As I've said a couple of times in the past -
our continued disablility to be able to advise people about implementing
DC in plain XML (rather than RDF/XML) using either DTDs or XML Schemas is
a big problem for us. Saying to people 'you made the wrong technology
choice' is simply *not* good enough. If we find we have to say that to
the majoritory of implementors then we made the wrong technology choice.
Some guidance (even in the form of notes) about implementing DC
(qualified and simple) in XML would be very helpful to people I would have
thought.
On the other hand, this is an Open eBook implementation of DC. Software
that processes this stuff will be Open eBook software - so it actually
doesn't matter too much what implementation choices they've made!
Andy
--
Distributed Systems and Services
UKOLN, University of Bath, Bath, BA2 7AY, UK [log in to unmask]
http://www.ukoln.ac.uk/ukoln/staff/a.powell Voice: +44 1225 323933
Resource Discovery Network http://www.rdn.ac.uk/ Fax: +44 1225 826838
|