Re: [office-metadata] summarizing recent suggestions

[Michael Brauer - Sun Germany - ham02 - Hamburg <Michael.Brauer@Sun.COM>]

Hi Elias,

Elias Torres wrote:
> Michael.Brauer@Sun.COM wrote on 02/28/2007 09:21:06 AM:
> 
>> Hi Bruce,
>>
>> Bruce D'Arcus wrote:
>>> On Feb 28, 2007, at 8:44 AM, Michael Brauer - Sun Germany - ham02 -
>>> Hamburg wrote:
>>>
>> My reply was confusing, sorry. I think what must have in the content is
>> enough information to efficiently update the field. Having some kind of
>> type (as suggested by you) and a link to the RDF/XML streams that
>> contains the real meta data (as suggested by me) should be sufficient.
>> Having just an id that is referenced from an arbitrary RDF/XML stream
>> probably isn't.
> 
> I'm fine with having a type and xml:id. A link to the RDF/XML is too
> constraining. Remember RDF is extensible and triples can come from multiple
> files to say that everything is inside one file, it's limiting in my
> opinion, but I'd compromise as long as it's a *hint* or starting point and
> not restricted to only one file.

Yes, it is just a hint. The type information could be used figure out 
which plug-in may provide the display string, and the plug-in would get 
the IRI of the RDF-XML stream. It could use it, or it could use hard 
coded file names. And it of cause could use additional streams. That's 
all up to the plug-in. So the IRI of the RDF/XML stream would, as you 
say, just provide a starting point, so that the plug-in does not have to 
search all RDF-XML streams to figure out where the data it is interested 
in is stored.

Michael