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

["Bruce D'Arcus" <bdarcus@gmail.com>]

On Oct 24, 2006, at 11:54 AM, patrick@durusau.net wrote:

> Bruce D'Arcus has the most complete proposals to date in the
> Metadata SC and we should do a quick overview to isolate issues  
> that require more
> detailed discussion.

Let me just flag some issues ahead of time:

1) I think the model and syntax is, from a broad view, easy, but
we're left with questions about details like:

a. should we constrain the model and syntax as I have (basically a
kind of RDF Lite), or not?

b. how do we manage the compatibility issue? Existing ODF metadata is
very close to this, but not quite (might be better to discuss later
with the TC as a whole, but am not sure)

2) If we go with the module approach I am advocating, how do we
formalize that? We could, for example, write RELAX NG schemas for the
modules (as I have), or not, and just use natural language
descriptions. We then just say so long as the syntax validates
against the generic schema, it's fine. I have no strong opinion here.

3) The most difficult bit, and what might be a candidate for partial
deferral to a later release, is the content association stuff;
including fine-grained statements within the content. This gets back
to our question "what do we want to describe in the document?" (which
I have not yet tried to answer).

4) Sort of related to 3, we also have the question of how to identify
documents and fragments within them for external reference. Could
also maybe be deferred, at least until the core is done.

Finally, I'm hoping Dan Brickley can help with some of these details,
but he'll be on a plane to Buenos Aires tomorrow (where he'll be for
the next six months), so won't be able to join us unfortunately.

Bruce