Re: [office-metadata] Groups - Metadata SC meeting added

["Florian Reuter" <freuter@novell.com>]

Hi,

my current understanding of meta:text-field is as follows.

We have two different forms. One is "the classical" using an xml:id

<text:meta-field xml:id="id1">
[Replacement text]
</text:meta-field>

in order to attach metadata to this field we use the <odf:Element rdf:About="IRI1" odf:path=".." odf:id="id1"/> linking
mechanism and make statements about IRI1 using RDF streams defines by <odf:Entry/>.

Second form is using an "inline" mechanism using the <text:meta-field-description> (this is my name for it; I guess we
have not yet a common name): 

<text:meta-field>
<text:meta-field-description office:process-content="false">
[RDF/XML]
</text:meta-field-description>
[Replacement text]
</text:meta-field>

Please note that the office:process-content="false" attribute is there for backward compatibility to ODF1.0/ODF1.1. It
prevents ODF reader to tread the content of <text:meta-field-description> as normal content.

Basically the <text:meta-field-descritpion> can contain anything an RDF/XML <rdf:Description> element can contain. 
The diffence to the first form is, that instead of a "named node" IRI1 an "anonymous node" is generated in the RDF
model.

===

Sample for case I:

content.xml:
<text:p>
<text:meta-field xml:id="id1">Hello World</text:meta-field>
</text:p>

with <odf:Element name="http://mysamples#Hello_World"; odf:path="content.xml" odf:id="id1"/> 

rdf.xml:
<rdf:Description rdf:about="http://mysamples#Hello_World";>
<dc:author>Florian</dc:author>
</rdf:Description>

will lead to the statement (http://mysamples#Hello_World, dc:author, "Florian") where http://mysamples#Hello_World
identifies the field with id1 in stream content.xml


Sample for case II:

content.xml:
<text:p>
<text:meta-field>
<text:meta-field-description office:process-content="false">
<dc:author>Florian</dc:author>
</text:meta-field-description>Hello World</text:meta-field>
</text:p>

will result in the statement (_:1 , dc:author, "Florian") where _:1 identifies the field above.


I also understand that text:meta-fields can be nested.

Is my understanding correct?

~Florian


>>> Bruce D'Arcus <bruce.darcus@OpenDocument.us> 04/04/07 2:29 AM >>>

On Apr 3, 2007, at 7:52 PM, Patrick Durusau wrote:

> I had gotten the impression that the RDF was doing to declare the 
> semantics of the field, that is a generic field mechanism with a rule 
> that whatever content is has is preserved on save so that a 
> non-enabled application can see the data.

No, we've had no resolution.

We spent months solving the question of in-content literals, but I 
think in-content resource references (e.g. text:meta-field) have gotten 
less careful attention. And I actually think they're more critical to 
this work.

I think Svante and Florian understand what I've been saying (in part as 
a result of some off-list discussion), but am not sure.

Bruce