Re: [ALL] RDF/A Primer Version

Ben Adida wrote:
> 
> 
> Continuing on this point, I guess that implies the following thing:
> 
> If http://example.com/foo resolves to an XHTML document, then 
> http://example.com/foo#bar can only be an information resource. However, 
> if http://example.com/foo resolves to an N3 document, then 
> http://example.com/foo#bar *can* be an information resource, because 
> there is no way to get a fragment of an N3 document.
> 
> The first consequence is that XHTML documents are somehow second-class 
> citizens to N3 in expressing semweb statements. That would be truly 
> unfortunate.
> 

It also seems to force a separation between the Web and SemWeb, which 
seems to me to be wholly against my understanding of the spirit of our 
activity.

So in my analysis of foaf identities of TBL, DanC and NormW in

http://lists.w3.org/Archives/Public/public-rdf-in-xhtml-tf/2006Jan/0065

TimBL is distinguished by having a URI (with fragment) where the URI 
(without fragment) only has an N3 and an RDF/XML representation. This 
avoids the problem we are discussing but at the cost of if you find 
Tim's URI for himself of:
http://www.w3.org/People/Berners-Lee/card#i
and stick it into a Web browser, then you get gibberish (unless you are 
SemWeb enabled).

I like the interwovenness of Norm's URI for himself of:
http://norman.walsh.name/knows/who#norman-walsh
which works both as a Web URI and a SemWeb URI, and I would assert that 
it works without any practical ambiguity. I see any claim that there is 
ambiguity between the HTML version of the URI and the RDF/XML version as 
largely dancing on pinheads, rather than any articulated 
interoperability failure. I, of course, enjoying dancing, particularly 
on pinheads ... but ... the goal to me is about making the SemWeb real 
by leveraging the success of the Web. This requires that the URI spaces 
for the Web and SemWeb are the same (or at least with large overlap), 
and not disjoint.

I feel it is very important that the SWBPD and the SemWeb community 
should resist an interpretation of WebArch that does not permit URIs to 
be used both as operational instructions on the Web for finding text, 
and as identifiers in SemWeb for arbitrary things. Such an 
interpretation will be a significant obstacle to the take up of SemWeb.

Jeremy


[[Note: in the apparent differences of opinion between me and David 
Booth, I hope it is clear that there is not yet an HP position, and each 
of us are primarily speaking personally, rather than in a representative 
capacity. I suspect that we will need to come to an 'HP' viewpoint 
sooner or later.]]

PS Perhaps the resolution of this will be that if you want a URI to be 
semweb enabled (at all) then your webserver should always give a 303 for 
it .... (oh how horrid, but at least implementable)

Received on Wednesday, 1 February 2006 12:23:13 UTC