Re: "role" and inheriting "about"

Yoshio Fukushige wrote:
> Hi RDFa people,
> 
> A question on RDFa "role" and inheriting "about."
> 
> Per RDFa Primer 1.0 (2006-05-16 version) [1],
> one can use "role" to introduce a new resource of a certain kind
> 
> e.g.
> 
> <p role="cal:Vevent">
>  ...
> </p>
> 
> (cf. 2.2 Publishing An Event)
> 

I believe role was added last minute to the primer but not to the syntax
spec and it's definitely either not final or underspecified. We were
discussing a couple meetings ago whether we would use @role or reuse
@class. We did not reach a decision at the time, but it is one of our
central issues right now.

> However, "4.2 Inheriting about" says
> 
> -------
> ... if an element carries a rel or property attribute, but no about attribute, 
> an RDFa browser will determine the subject of the RDF statement by navigating up 
> the parent hierarchy of that element until it finds an about, 
> or until it gets to the root element, at which point the default is about="".
> -------
> 
> My question is what if an RDFa browser finds a "role" attribute on its way 
> climbing the tree.

[non-normative] Right now my Python parser follows the subject
resolution mechanism ignoring the presence of a role attribute in the
three[/non-normative]

> 
> e.g.
> <p role="cal:Vevent">
>   I'm giving 
>   <span property="cal:summary"> a talk at the XTech conference </span>
> </p>
> 
> Does it stop climbing and set the resource introduced by the "role" attribute 
> as the subject?
> 
> e.g. does it yield
> 
> [a cal:Vevent; cal:summary "a talk at the XTech conference"].

>From the primer, I'd assume that this is the intended case. I have
limited background on this because I joined very recently but there are
no other subjects in the full-HTML example and the section showing the
triples generated doesn't use <> as the subject.

> 
> ?
> 
> Or a "role"  subject is described only by "meta property" expression 
> to which the "inheriting about" rules does not apply?
> 
> , which will yield
> 
> <> cal:summary "a talk at the XTech conference".
> 
> ?

There's some inheriting when using meta properties, especially depending
 on the location of the meta element. If in head, it's the document, but
elsewhere it's either the parent's about or a blank node.

> 
> To me, it sounds more natural for the search to stop
> when a "role" is found on its way.

It feels like a good thing to do, but I'm not sure is right, if anything
it needs more discussion in my opinion. Let me explain what I mean:

<p>
  <link rel="rdf:type" href="cal:Vevent"/>
  <span property="cal:summary">...</span>
</p>

What should the subject of cal:summary be? If we say to stop at role,
then should we stop at p's (bnode) as well. The question is whether an
author adding an rdf:type triple implies it also implies a new subject
or not.

> 
> Do I miss or misunderstand something? 
> (or misdeduce?)

Not at all Yoshio, if anything we need to do a better job of being more
explicit, but it's all a work in progress and thanks for your
questions/feedback.

> 
> [1] http://www.w3.org/TR/2006/WD-xhtml-rdfa-primer-20060516/
> 
> Best,
> Yoshio Fukushige
> fukushige.yoshio@jp.panasonic.com
> 
> 

Elias Torres

Received on Friday, 8 September 2006 17:40:05 UTC