Re: [Fwd: Re: [office-metadata] Reuse of metadata proposal for non ODFapplications]

[Svante Schubert <Svante.Schubert@Sun.COM>]

Bruce,

Bruce D'Arcus wrote:
> Svante Schubert wrote:
>
>> Hmm.. When I see it, would someone expect that the text:meta-field 
>> element as part of a RDF schema would have the same features as in an 
>> ODF file?
>> I don't think so, why giving things many names, when they are the 
>> same in the end.
>
> As I said on the call, that class is not formally defined anywhere. 
> It's just some convention you invented. 
People on this list might have problems to follow our discussion.

As background information, I was referring to an earlier discussion
http://lists.oasis-open.org/archives/office-metadata/200706/msg00040.html
and wrote that instead inventing a OWL class like odf:field for 
text:meta-field, we might as well use a generic mechanism to identify 
certain elements from ODF.

Is this the problem / the requirement we are talking about? The 
Identifying the text:meta-field in the user metadata RDF/XML?

In assumption of that I proposed to introduce (in ODF 1.3) a general 
naming mechanism instead of inventing names for certain ODF elements.
For example there might be metadata extension working on table:table, 
possibly even creating or adapting them.
In this case the table element would be already identified to be an ODF 
element, using the OWL class describing odf:Element, but might be as 
well be of the RDF type table:table.

This would be written as:

<odf:Element rdf:about="uri:someIRI" idref="someID">
  <rdf:type 
rdf:resource="urn:oasis:names:tc:opendocument:xmlns:table:1.0table"/>
</odf:Element>

The text:meta-field could be identified similar as the following:

<odf:Element rdf:about="uri:someIRI" idref="someID">
  <rdf:type 
rdf:resource="urn:oasis:names:tc:opendocument:xmlns:text:1.0meta-field"/>
</odf:Element>

or - as the type could be used as RDF element name - written as

<text:meta-field rdf:about="uri:someIRI" idref="someID">
  <rdf:type 
rdf:resource="http://docs.oasis-open.org/opendocument/meta/package/odf#Element"/> 

</text:meta-field>

> From an RDF perspective, it's pretty much useless.
This surprises me, as it fullfils the requirement to identify an ODF 
element to be a text:meta-field.
Did I misjudged your requirements?
>
> Also as I said, fields are pretty critical semantically. In some other 
> file formats, they might be represented in an attribute with content 
> like "\d \s doe999 \t". In ODF 1.2, we do it in RDF/XML. We're really 
> not going to say anything about how to do that??
All the citation extension needs to know is where the citation field is 
in the content, this is done by the xml:id on the field. The 
identification of being a citation is done by the 
rdf:type="uri:someCitationUri" in the metadata manifest. This manifest 
includes as well a reference to the related user RDF/XML. This user 
RDF/XML is finally free to use any RDF they desire to use.
>
> What I am suggesting, then, is formally defining a field in OWL:
>
> odf:Field a owl:Class .
>
> .. which then allows me to subclass that:
>
> odf:Citation a owl:Class ;
>     rdfs:subclassOf odf:Field .
>
> Likewise, I have pointed out that generic properties for the field 
> like prefix and suffix are perfectly in order.
I would have assume that the prefix and suffix would be part of the user 
RDF/XML file.
>
>> I am not sure if this really have to become part of 1.2. I would 
>> rather let it be, as our timeline has finally come.
>
> With all due respect, I've waited three years to have something in the 
> spec to handle citations. You're now asking me to simply forget all of 
> that, even the most generic things, and wait another few years???
I asked for a reason to break the time line. It seemed more adequate to 
me to mature the citation vocabulary during implementing a prototype 
instead giving limits to the citation RDF vocabulary in half a week.
>
> This goes to my question --- which I've been repeatedly asking for 
> months -- about how we're going to encourage common implementations 
> going forward. I'm not set on it being in the spec now, but I do 
> insist that we do *something* to address this.
I completely agree on encouraging implementation by further assistance 
from the metadata SC, just how the odf:Field could encoure 
implementations is not comprehensible to me.

regards,
Svante