Re: named graphs (was Re: Data URIs)

On 2 Jul 2008, at 20:32, Sandro Hawke wrote:
[snip]
> To rephrase: let's make sure our approach does not preclude a  
> design for
> named graphs which works smoothly with OWL 2.
>
> We can do that, more or less, with an existence proof: come up with a
> (non-standard) design for named graphs which works smoothly with  
> OWL 2.

Wow, this is a pretty high bar.

> Of course some future WG might not pick that design, due to other
> constraints, but I think we'll at least have done our part.


We're going to hand them a fair bit of legacy to deal with.

Let me go back to what you wrote earlier:

> As far as I can tell, there is consensus in the RDF community that  
> there

> should be a named-graph standard.  Making such a standard is,

> unfortunately, an awkwardly-sized peice of work.  It doesn't really  
> fit

> inside any other work, and it's rather small for a whole new Working

> group.  So I don't know exactly when/how it will happen.

I presume you don't mean "named graphs" per se, but some reasonable  
context/quotation mechanism.

It seems to me that we have a couple of choices. Unfortunately, I  
think punting on any feature that needs that is simply a non-starter.  
So:

1) We could allow it in functional and XML, but not in the RDF since  
we'd wait for the future work for that.
	Pro: Life is easier for us. No forward compatibility worries.
	Con: Forces people to use the XML format for some features.

2) We do our own thing with some claim about planned obsolescence.
	Pro: That's what we're doing now! Some level of compatibility.
	Con: Our task is harder because we can't dork with RDF syntax or  
tools, really. Doing a ton of work for a hack that is intended to be  
superceded seems...wrong.

3) We spec a real solution.
	Pro: Solves the problem in what we hope is a right way. Solves the  
awkward sized work issue. Gets them out there at last.
	Con: We're not chartered for that. Requires finding more  
stakeholders to participate.

3.1) W3C specs a new solution. That is, sorta like with the  
international string thing, we spin off a minigroup to do this. We  
solve our problem in XML and that becomes the guiding rationale of  
the group (which helps scope the problem cleanly). We defer to that  
group, potentially holding the RDF mapping back until its done but  
let the other parts move ahead.

We have a need and it's not filled. Our need is concrete. This  
wouldn't be the first time that the needs of OWL drove some rdf stuff :)

I recognize that 1, 3, and 3.1 are...bold...proposals. But the only  
thing stopping them is lack of will, I'd guess. But how hard would it  
be to get a lil XG together of RDF folks with a liason from us and  
say, "fix this; we can rec it in OWL WG in the worst case".

Cheers,
Bijan.

Received on Friday, 4 July 2008 13:46:16 UTC