Re: skos and dc example

[Bruce D'Arcus <bdarcus@gmail.com>]

cc-ing the ODF metadata list since Mikael has some suggestion for Svante 
and Patrick ...

Mikael Nilsson wrote:

> Nice! I'm very happy to see that ODF gets such an excellent metadata
> framework. Do you have any recommendations on vocabulary?

Good question!

No, but this is why I was asking the questions about the new dcterms 
stuff. E.g. I would like us to suggest vocabularies, even if informally. 
I think it would be bad form to just introduce this system and not give 
developers help in figuring out how to most effectively use it.

So I would obviously like to suggest DC, and preferably that people 
prefer to use the new dcterms properties.

In fact, in the next couple of weeks I need to provide a mapping from 
the old BibTeX key-value citation support to RDF so that we can move 
citation support to the new framework. I'm working on that as part of 
the work at bibliontology.com, but we're planning to use as much of DC 
as we can, and we're faced with the problem: do we use dc:date or 
dcterms:date, dc:title or dcterms:title?

Note: the citation stuff is independent of ODF, but we'll be faced with 
the same problems with ODF more generally.

You once offered to liason with us Mikael. We'd happy if you have any 
suggestions on this issue going forward.

My preference would be that we issue a kind of 
best-practices/tutorial/suggested vocabularies document or set of 
documents; probably at the URI for the metadata namespace. I'd like if 
we could suggest using the new dcterms properties as soon as the DCMI 
has finalized them.

Am also hoping to see the new vCard-in-RDF stuff stabilize so we can 
suggest to use it.

> A few (surely late, but still) comments:
> 
> 1. The indentation in the ontology snippets makes it unnecessarily hard
> to read.

Agreed. XML fragments in the spec ought to be indented.

> 2. Section 1.2 on in-content metadata seems to duplicate a lot of the
> work on RDFa  ( http://www.w3.org/TR/xhtml-rdfa-primer/ ), which is
> focused on XHTML, but still. 

Yes, that is by design. Elias Torres from IBM (also very involved in 
RDFa) is responsible for most of that. We did adapt some of it obviously.

We tried to reuse as much as possible from elsewhere; the only thing I 
think we did that was really new was a) defining a mechanism to bind 
external RDF statements to XML nodes (which, come to think of it, could 
be used in XHTML), and b) the new RDF-based manifest for the package.

Oh, and the new generic field describe in RDF is pretty clever if I do 
say so myself ;-)

>>> Regarding your question about the "unnecessary" references to the DCAM,
>>> it *is* true that RDF implementors will just need the RDF schemas, plus
>>> human-readable term descriptions. 
>> But I'm not sure that's clear for the average developer, who may have no 
>> background with the DCMI work, and may be new to RDF.
> 
> Right. I think we need some tutorial-level material in this domain...

That'd be great.

> Note that all these specs are pretty new in their current incarnation,
> so this is in the pipeline....
> 
>>> They have not yet been finalised, but are in progress. You'll be able to find them at
>>>
>>> http://dublincore.org/documents/dcmi-terms/
>> Thanks. Any ETA?
> 
> Well, it's in public comment currently, but Tom might be better
> positioned to tell when they might go "live". but certainly within a few
> months, we all hope.

OK, thanks.

Bruce