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

[Bruce D'Arcus <bruce.darcus@OpenDocument.us>]

On Sep 5, 2006, at 7:18 PM, patrick@durusau.net wrote:

> 2. Collaborative Authoring Requirements
> 3. Enhanced Search Requirements

Just a reminder that I moved the Writely doc to here:

<http://wiki.oasis-open.org/office/ 
OpenDocumet_Metadata_Use_Cases_and_Requirements>

Perhaps given all our own "collaborative authoring" work, we might  
start again by discussing what kind of metadata might be important to  
this use case, and whether the above requirements cover those needs?  
I've not really gone back to Rob's document to see if I'm missing  
anything.

In one of John Barstow's earlier notes, he mentioned WRT to this use  
case something like "the open world assumption is your friend", by  
which he meant that by default RDF understands the statements contained  
in a document/graph as open; that there may be other descriptions of  
the same content elsewhere, and that they can be merged.

So using a compact RDF syntax, say you have three statements about the  
same document, identified by the URI  
"http://wiki.oasis-open.org/office/ 
OpenDocumet_Metadata_Use_Cases_and_Requirements".

<http://wiki.oasis-open.org/office/ 
OpenDocumet_Metadata_Use_Cases_and_Requirements> a meta:Document;
	dc:title					"OpenDocumet Metadata Use Cases and Requirements" .

<http://wiki.oasis-open.org/office/ 
OpenDocumet_Metadata_Use_Cases_and_Requirements> ;
	dcterms:contributor		"Bruce D'Arcus" .

<http://wiki.oasis-open.org/office/ 
OpenDocumet_Metadata_Use_Cases_and_Requirements> ;
	dcterms:contributor		"Patrick Durusau" .

So, three discrete statements about the same thing, which could be  
stored in different locations.

But because of the model, if you run these through an RDF parser, it  
will automatically merge the statements into a single description  
basically (using the URI as key).

I think this open world notion and ability to merge statements is  
fundamental to this case, in fact (and maybe to "extrinsic metadata").

 From an XML standpoint, it means that properties -- as Rob pointed out  
-- have to be represented as elements, so that the above merged  
description would be:

<meta:Document  
rdf:about="http://wiki.oasis-open.org/office/ 
OpenDocumet_Metadata_Use_Cases_and_Requirements">
   <dc:title>OpenDocumet Metadata Use Cases and Requirements</dc:title>
   <dcterms:contributor>Bruce D'Arcus</dcterms:contributor>
   <dcterms:contributor>Patrick Durusau</dcterms:contributor>
</meta:Document>

This also ties to my interest in allowing description to link to other  
descriptions and the "automatically generated metadata" case.

For example, imagine a collaborative editing scenario as above, but  
where the editor in question has full contact information for the  
contributors. Instead of using dumb strings to represent the  
contributors, it could instead automatically embed a vCard  
representation of them with full contact info, and link to them as full  
resources/objects.

<http://wiki.oasis-open.org/office/ 
OpenDocumet_Metadata_Use_Cases_and_Requirements> ;
	dcterms:contributor		<mailto:doej@ex.net> .

<mailto:doej@ex.net> a v:VCard ;
	v:fn						"Jane Doe" ;
	v:title					"CTO" ;
	v:org	[
		v:orgname				"ABC, Inc" .
	] .

Bruce