Re: ISSUE: Using @id to set subject in RDFa (ISSUE-121)

Sebastian,

(This is my personal answer, and not an official answer from the group)

this is not an error report on an existing feature; instead, it is a request for a new feature. As such, my first question is: is there a convincing evidence that this functionality is needed by the community?

In my personal experience as a user, but also getting feedbacks from other users, I do not see this as a major feature request from the community. At least in my case, almost all (if not all, I have not checked) usage of @about uses a URI that does _not_ bind to an @id value (whether that URI uses a fragment id or not). The reason is that when encoding information in RDF on some specific subject, that subject's URI is usually an 'abstract' URI (URI of a person, a paper, a presentation, etc) which is not directly represented in the HTML source via an element id. An @id may be used, though, as part of the object of a triple, ie, via @href or @resource (bound to predicates like 'described by', or similar).

That is at least my experience. Based on that I am not in favour adding yet another feature to RDFa which is already quite rich (other would say 'complicated') in features...

Sorry about that...

Cheers

Ivan


On Nov 19, 2011, at 19:02 , Sebastian Heath wrote:

> Currently, RDFa processors do not take account of @id attributes in
> host-languages. In particular, in (x)html the @id attribute is
> ignored. In (x)html, this creates unnecessary complication in which
> the value in @id needs to be prefixed by '#' and put in an @about
> attribute in order to make semantic markup visible to both html agents
> and rdfa processors.
> 
> Example:
> 
> <p id="item1" typeof="ex:item" about="#item1">
>   <span property="item_name">An interesting item (1)</span>
> </p>
> 
> I suggest that a combination of @typeof and @id cause the subject to
> be set to the fully qualified URL implied by the value @id, according
> to normal URL processing rules defined for HTML and related languages.
> 
> Given a document http://example.org/document1 with appropriate
> @vocab/@prefix definitions, this would lead to the markup
> 
>  <p id="item1" typeof="ex:item">
>   <span property="rdfs:label">An interesting item (1)</span>
> </p>
> 
> producing the triples
> 
> <http://example.org/document1#item1> rdf:type <http://example.org/ns/item> .
> <http://example.org/document1#item1> rdfs:label "An interesting item (1)" .
> 
> The main advantage is the simplicity of using a single construct to
> make semantic data visible to both browsers, DOM aware languages such
> as javascript, and to RDFa processors.
> 
> The restriction that the subject is only set when there is also a
> @typeof will reduce the number of so-called 'junk' triples. But I note
> that RDFa accepts the generation of such triples in other situations
> such as link elements invoking style sheets using @rel.
> 
> It may be that this functionality is best defined for particular host
> languages and not in the RDFa 1.1 core.
> 
> Sebastian Heath.
> 


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

Received on Sunday, 20 November 2011 08:39:06 UTC