Re: [office-metadata] Content duplication and ODF related RDF vocabulary

[Elias Torres <eliast@us.ibm.com>]

Svante.Schubert@Sun.COM wrote on 12/18/2006 09:09:48 PM:

> Hi Elias,
>
> Sorry for the confusion, I will restate my hypothetical example.
>
> Elias Torres wrote:
> > Below are my responses. BTW, I think people might be thinking that it's
> > either RDFa or xml:id and I want to reiterate that it's not the case.
RDFa
> > is a superset of xml:id. RDFa *does* use xml:id to identify resources
> > withing the content.xml. Supporting RDFa doesn't meant abandoning
xml:id,
> > it simply means allowing us to avoid content duplication and more
> > in-context metadata.
> >
> Thanks for clarifying this.
> > Svante.Schubert@Sun.COM wrote on 12/18/2006 07:54:41 PM:
> >
> >
[snip]
> >
> >
> My example is like this:
> There are four span with different styles, which have a word as content
> from fooA to fooD.
> One semantic is for all of them, "http://ex.net/foo1";
> And then there are two splitting and one intersection of semantic:
>
> The semantic 'http://ex.net/foo2' is related with the content 'fooA'and
'fooB'
> The semantic 'http://ex.net/foo3' is related with the content 'fooB'and
'fooC'
> The semantic 'http://ex.net/foo4' is related with the content 'fooC'and
'fooD'

This is the kind of question I was trying to avoid with my email on the
usage of xml:id and RDFa. First of all, content can't be referenced in RDF.
Let me give you an example.

<#elias> :age "62" .

How can somebody refer to my age? The straight-forward way is to use
duplicate the content.

<#eliasTwinBrother> :age "62" .

However there's no relation between one "62" and the other.

In order to reference it has to be modeled as a resource so others can
reference it.

<#eliasAge> rdf:type age:Age ;
         rdf:value "62" .

Now somebody can refer to my age.

<#eliasTwinBrother> :age <#eliasAge> .

Again, the example is definitely stretching things a bit but I'm trying to
think what the use-case scenario for that is. In other words, I'm trying to
think how an application would come up with such a model that 4 different
things want to re-use the literal value of one of them with different
combinations and all of this by reference with no content duplication.
Would you please show me how does xml:id-alone or styles solves this
problem? Once you see how complicated this in RDF/XML alone, I would then
show you how can some of those triples would end up in the metadata (if it
makes sense at all, since remember I'm not saying that EVERYTHING must be
in the content.xml).

-Elias