Default context setups... (fwd Questions on the Link Registry for RDFa (ACTION-100))

Dear all, 

in view of Mark's answers below, I propose the following:

- We should use the http://www.w3.org/1999/xhtml/vocab#blabla for the link term URIs. It seems that IANA does not really care, and there is therefore no reason to break backward compatibility for all the terms that are common in this list and the older one. Note, for example, that the Creative Common spec contains a a sameAs between cc:license and the xhtml/vocab#license, which is a very good reason _not_ to change the URI-s

There is one exception, and that is the described by coming from Powder. Niklas proposed to add that to the same base URI and set up an equivalence to the powder URI. That of course is theoretically nice, but I do not expect RDFa processors to follow their nose on the default context. Ie, I would prefer to keep the powder URI.

- The http://www.iana.org/assignments/link-relations/link-relations.xml seems to be the right URI for the list. I am not sure this list appears in the Rec, though, it should appear only in the default context document.

- We *could* go out of our way to get a http://www.w3.org/context/ top level URI for the default context document. However... I am not sure we would get it ('context' is a very general term). Also, even if the term context appears in JSON-LD, the way things evolve is that the terminology and vocabulary used in a JSON-LD file is different than for RDFa; trying to get a common top level URI would only muddle the waters. Finally, now that we do not have RDFa profiles, these really would be just for documentation, we expect implementation to hardwire the content. 

Bottom line, I would propose to make it simple and not complicate our lives, and use the following two URI-s

http://www.w3.org/2011/rdfa-context/rdfa-1.1
http://www.w3.org/2011/rdfa-context/html-rdfa-1.1

I will implement these, we can always move them if you guys have a strong feeling about this and want me to change.

Any thoughts on this?

Ivan




Begin forwarded message:

> Resent-From: public-rdfa-wg@w3.org
> From: Mark Nottingham <mnot@mnot.net>
> Date: October 21, 2011 01:45:56 GMT+02:00
> To: Ivan Herman <ivan@w3.org>
> Cc: Manu Sporny <msporny@digitalbazaar.com>, W3C RDFWA WG <public-rdfa-wg@w3.org>
> Subject: Re: Questions on the Link Registry for RDFa (ACTION-100)
> archived-at: <http://www.w3.org/mid/685EC6AB-B220-4EBB-AF7A-B6DC1CCC516E@mnot.net>
> 
> Hi Ivan,
> 
> On 21/10/2011, at 2:42 AM, Ivan Herman wrote:
> 
>> Mark,
>> 
>> as you may know, RDFa 1.1 is in its finishing round (the editors' draft is available at [1]). One of the last issues that the group has to decide is how to interpret, when generating RDF, a statement like
>> 
>> <a rel="author" href="blabla">....</a>
>> 
>> ie, what RDF predicates should be generated (if any) for the value of @rel. RDFa 1.0 used the rel relations as defined in the XHTML document, and generated a predicate in the xhtml/vocab namespace, e.g.,
>> 
>> http://www.w3.org/1999/xhtml/vocab#author
>> 
>> for all @rel values that XHTML1 defined. For RDFa 1.1, the situation is a little bit complicated, because it is unclear what @rel relations the HTML WG will define for HTML5[2]. It will be different, probably, than the old list.
>> 
>> On our meeting today, the RDFa WG decided that, instead of any of the HTML specs, we should use the IANA link relation list:
>> 
>> http://www.iana.org/assignments/link-relations/link-relations.xml
>> 
>> which seems to be much more complete than what is in the current HTML document[2] anyway.
>> 
>> However, there are some details that we should clarify, hence this mail. These are:
>> 
>> 1. What is the authoritative URI to refer to for this list? Is the URI above all right? We must admit we simply got there via google:-)
> 
> Yes. When we do 5988bis, it should include the URL.
> 
> 
>> 2. Does IANA have any advice/requirement as for the URI-s to be used when materializing those link relations in RDF? As I said, we used the xhtml/vocab namespace before, and we can certainly continue doing that for most, although the describedby relation already has a URI in the Powder namespace. Or, alternatively, do you guys have fixed URI-s that you'd prefer us to use?
> 
> Nothing fixed; probably the IANA URIs make sense, but you could certainly define something else as a base.
> 
> 
>> 3. How frequently do you plan to update this list? 
> 
> It's a living document; as we get requests, we'll update it. 
> 
> Note also that there's an effort underway to make web-related IANA registries easier to use [1]. This should improve its coverage, over time. It also might be good to ask this question there, and/or on the link-relations list.
> 
> Cheers,
> 
> 
> 1. http://www.w3.org/wiki/FriendlyRegistries
> 
> 
> --
> Mark Nottingham   http://www.mnot.net/
> 
> 
> 
> 


----
Ivan Herman, W3C Semantic Web Activity Lead
Home: http://www.w3.org/People/Ivan/
mobile: +31-641044153
PGP Key: http://www.ivan-herman.net/pgpkey.html
FOAF: http://www.ivan-herman.net/foaf.rdf