Re: [office-metadata] Binding proposal

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

On Mar 19, 2007, at 12:58 PM, Patrick Durusau wrote:

> Svante Schubert wrote:
>
>> Hello group!
>>
>> After several discussions here at Sun and a week-end of 
>> reconsideration, one thing seems clear now:
>> It makes little sense to trust on an identification mechanism, that 
>> can be so easily annulled as by a file copy.
>> We doubt that any company would rely their data on such fragile 
>> mechanism, unless a special work-flow is being used for documents in 
>> the company.
>>
> Sorry, that went by a little fast. What do you mean the identification 
> mechanism is "annuled as by a file copy."?

The problem issue here is the situation where the subject or (URI) 
object of a statement is an ODF content node, as identified by a 
(local) xml:id.  The URI, then, has to be constructed to be able to 
reference that node.

Likewise, if we want to treat a document as a subject.

> I attach metadata in an ODF document. I then copy the document. How 
> has that "annuled" my metadata?

Let's take the simple case of a document with a single paragraph, 
identified by the xml:id value of "foo". We want to make statements 
about that paragraph.

We need to know what to put as the value for the RDF/XML rdf:about 
attribute. We cannot do:

<rdf:Description rdf:about="content.xml#foo">
...
</rdf:Description>

... because we cannot use a relative URI here. We have to establish the 
base of the URI.

So say we decide the base URI should be a UUID encoded as a URN. We can 
then do:

<rdf:Description rdf:about="urn:uuid:[uuid]/content.xml#foo">
...
</rdf:Description>

The problem is that if somebody does a cp on the file system, we now 
have two documents with the same identifier. We don't yet have a 
problem if the content is the same, I think, but what happens if the 
user opens the second document and adds a second paragraph, and adds 
statements to that?

> Actually I was thinking that metadata would continue to be valid in a 
> document until someone changed it. Which would mean that documents 
> would get "richer" with metadata as they are process in a work flow.
>
> The mechanism isn't "fragile" at all but robust and therefore 
> inconsistent with processing models meant for truly "fragile" metadata 
> as we have today.

Right. I think Elias' solution to the above is to do something like:

<rdf:Description 
rdf:about="tag:odf:[uuid-base]:[uuid-version]/content.xml#foo">
...
</rdf:Description>

*That* would solve the CP case, because upon saving the edited copy, it 
gets a new version id.

Bruce