OASIS Mailing List ArchivesView the OASIS mailing list archive below
or browse/search using MarkMail.

 


Help: OASIS Mailing Lists Help | MarkMail Help

office-metadata message

[Date Prev] | [Thread Prev] | [Thread Next] | [Date Next] -- [Date Index] | [Thread Index] | [List Home]


Subject: Re: [office-metadata] Content duplication and ODF related RDFvocabulary


Bruce D'Arcus wrote:
>
> On Dec 19, 2006, at 8:44 PM, Svante Schubert wrote:
>
>> Let me write the content.xml as I assume it would like it in an 
>> xml:id-alone approach.
>> The content.xml using the xml:id-only approach, as Model independent 
>> View without awareness of any metadata
>>
>> <!-- some text metadata refer on from the content.xml -->
>> <text:span style:name="s2" xml:id="_1">CG7217</text:span>
>> <text:span style:name="s2" xml:id="_2"> is a gene product. </text:span>
>> <text:span style:name="s2" xml:id="_3">It was introduced on </text:span>
>> <text:span style:name="s2" xml:id="_4">GENCON Conference</text:span>
>> <text:span style:name="s2" >.</text:span>
>
> The RDF could be:
>
> _:some_anonymous_id a ex:GeneProduct ;
>     rdf:value "CG7217" ;
>     ex:introductionVenue "GENCON Conference" .
>
> So three triples. [note: the event could also be a resource itself]
>
> Options I would be fine with:
>
> 1)
>
> RDFa approach, with triples in-content.
>
> <text:p meta:about="http://ex.net/1"; meta:class="ex:GeneProduct">
>   <text:span meta:property="rdf:value">CG7217</text:span>
>   is a gene product.
>   <text:span style:name="s2">It was introduced on </text:span>
>   <text:span style:name="s2" 
> meta:property="ex:introductionVenue">GENCON Conference</text:span>
> </text:p>

>
> 2)
>
> RDFa approach, with triples in-package.
>
> <text:p meta:about="http://ex.net/1";>
>   <text:span>CG7217</text:span>
>   is a gene product.
>   <text:span style:name="s2">It was introduced on </text:span>
>   <text:span style:name="s2">GENCON Conference</text:span>
> </text:p>
>
> All that content gets duplicated in RDF/XML, and properties do not get 
> marked as such in content:
>
> <ex:GeneProduct rdf:about="http://ex.net/1";>
>   <rdf:value>CG7217</rdf:value>
>   <ex:introductionVenue>GENCON Conference</ex:introductionVenue>
> </ex:GeneProduct>
>
> Using xml:id to identify the subject would work, too, but they are 
> different.
>
> 3)
>
> RDFa approach, using a style abstraction:
>
> <text:p meta:about="http://ex.net/1"; style:name="Gene Product">
>   <text:span style:name="value">CG7217</text:span>
>   is a gene product.
>   <text:span style:name="s2">It was introduced on </text:span>
>   <text:span style:name="introductionVenue">GENCON Conference</text:span>
> </text:p>
>
> ... where the styles contain the appropriate property and class URIs 
> instead.
>
> 1 and 3 are essentially equivalent.
>
> I dislike the xml:id approach to achieve those outcomes because it 
> adds too much complexity while leaving the content pretty much opaque. 
> I would prefer to scale back functionality than do this.
>
How does the metadata 'know' to ignore the ODF XML elements?
There are meta:about containing text:span, how would the default RDF 
application react on this?

Imagine further the model for these parts are in the package. One is 
taken from the example of the RDF primer.
http://www.w3.org/TR/2004/REC-rdf-primer-20040210/#example43
Some real world model data. Do you have to adapt them? The xml:id 
approach not, before the content is changed by the user.
And do you think it is too complex to leave the model as it is and the 
content.xml like this:

text:span style:name="s2" xml:id="_1">CG7217</text:span>
<text:span style:name="s2" xml:id="_2"> is a gene product. </text:span>
<text:span style:name="s2" xml:id="_3">It was introduced on </text:span>
<text:span style:name="s2" xml:id="_4">GENCON Conference</text:span>
<text:span style:name="s2" >.</text:span>

Summarizing the relation between model and content is taking over the 
compexity, but it is simple in N3 similar to:

# The mapping relations expressed in N3 
(http://www.w3.org/2000/10/swap/Primer.html)
# The metadata refer to the RDF primer example 43 
(http://www.w3.org/TR/2004/REC-rdf-primer-20040210/#example43)
# the others are assumed to be in a file called xmeta.xml but not 
furhter listed.
#
# NOTE:
#   All ODF depended naming is just made up for this example and should 
be overworked.

@prefix loc:    <../meta/locations.xml#>.
@prefix sig:    <../meta/significance.xml#>.
@prefix gen:    <../meta/example43.xml#>.
@prefix edit:   <../meta/editors.xml#>.
@prefix desc:   <../meta/desc.xml#>.
@prefix con:    <../content.xml#>.
@prefix go:     <http://www.geneontology.org/xml-dtd/go.dtd#>.
@prefix odf:    <urn:oasis:names:tc:opendocument:xmlns:office:owl#>.
@prefix rdf:    <http://www.w3.org/1999/02/22-rdf-syntax-ns#> .

#
# ODF PROPERTIES FOR RELATION OF CONTENT AND METADATA
#
# overall base relation
odf:isRelated   a   rdf:Property.


#
# RELATIONS BETWEEN CONTENT AND METADATA

# 1)
desc:genprod    odf:isRelated     con:_1, con:_2 .

# 2)
desc:genIntro   odf:isRelated     con:_3, con:_4 .

# 3)
edit:frank      odf:isRelated     con:_2, con:_3 .

# 4)
sig:important   odf:isRelated     con:_1, con:_2, con:_3, con:_4 .

# 5)
gen:GO_0003673.go:association.go:gene_product.go:name
                odf:isRelate      con:_1 .

# 6)
loc:conf        odf:isRelated     con:_4 .



Athough instead of N3 path a XPointer to the Model would be more 
appropriate.

Regards,
Svante


[Date Prev] | [Thread Prev] | [Thread Next] | [Date Next] -- [Date Index] | [Thread Index] | [List Home]