process

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

I think the time for talk has virtually come to an end for this SC.

The conversation thus far has gotten us to the following point:

1) We agree on the majority of points

	- we support RDF/XML
	- we allow multiple RDF/XML files, registered appropriately through 
the manifest
	- we will suggest some conventions for serializing and maybe storing 
the RDF/XML
	- we want a way to add fine-grained metadata to the content

I think this consensus is clearly a good thing, as it is the majority 
of what we need to do.

Moreover, I think we have a good set of use cases and requirements to 
guide the rest of our design process.

2) We disagree on the final major requirement: "associating metadata 
with content" (see, I told you all this would be the most difficult 
:-))

	- Svante and Patrick are arguing for an xml:id-only approach
	- Elias and I for the hybrid in-content approach

Given the tone of the list discussion, I think we might need to assume 
that we have two sub-proposals, and perhaps stop worrying about trying 
to achieve some general consensus.

Given that the outcome of this process will ultimately depend on the 
discussion within the TC proper, can we perhaps just plan on delivering 
them:

a. a single document that reflects the consensus in 1
b. two documents that reflect the alternatives in 2

...?

That way we isolate the specific divergence from consensus, and we can 
get on with our work without endless conversation. And if one of those 
sub-proposals drops away before it gets to the TC (because we do find 
some consensus), that's fine too.

Just a thought ...

Bruce