Multiple content nodes representing on RDF subject

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

Hi group,

I would like to give you an example, how we might solve one specific 
aspect for metadata in ODF.

In this scenario there are multiple portions of content representing one 
RDF object.

Just imagine there is a certain text with semantic in the content:

<text:p xml:id="_foo1">This is an important and divvikult text!</text:p>

which is part of a N3 statement:

<#_foo1> <#isImportant> "This is an important and divvikult text!"

Now a user recognizes an error in the text and marks a subset with a 
different vocabulary.

<text:p xml:id="_foo1">This is an important and </text:p>
<text:p xml:id="_foo2">divvikult</text:p>
<text:p xml:id="_foo3"> text!</text:p>

By this a new RDF statement has been added into the existing text

<#_foo2> <#typingError> "divvikult"

Furthermore the earlier N3 statement has changed as well, as the sum of 
(<#_foo1> + <#_foo2> + <#_foo3>) representing now the earlier important 
text.

(Not sure if it is correctly written as a List in N3, but if so - ) it 
might be written as
(<#_foo1>,<#_foo2>,<#_foo3>) <#isImportant> "This is an important and 
divvikult text!"


One solution to satisfy this requirement is to create an indirection 
between content and metadata, which will be part of ODF XML 
implementation. This indirection bundles the various IDs - possibly from 
all over the package - representing one subject, with one ID.

<content xml:id="_foo">
  <ref idref="_foo1"/>
  <ref idref="_foo2"/>
  <ref idref="_foo3"/>
</content>
<content xml:id="_blah">
  <ref idref="_foo2"/>
</content>

The metadata might now reference to the ID given by this indirection and 
the ODF application would be able to identify
<#_foo>  <#isImportant> "This is an important and divvikult text!"
<#_blah> <#typingError> "divvikult"

Hopefully it gives you an impression about this certain problem and the 
way we would be able to solve it.


Hope all of you have a nice x-mas and wonderful presents!
Svante