Agenda, previous 2007-07-19
See also: IRC log
<scribe> post-meeting note:
ACTION: MichaelH to put together test cases for @resource
@href on same element [recorded in http://www.w3.org/2007/07/19-rdfa-minutes.html#action09]
[DONE]
-- see mail
from Michael
<markbirbeck> it will be a low turnout today!!
Ben: none of these are resolutions, but we have agreement that we move forward with implementations and test cases in order to resolve the remaining issues.
for now, we'll go with @instanceof
<benadida> span resource="#foo">
<benadida> </span>
Ben: we will try out striping where @resource, @href, and the bnode on a @rel or @rev becomes the subject of contained elements
<markbirbeck> for the purposes of explaining this to 'novices', we used to call this 'chaining'.
<benadida> <div id="foo" instanceof="foaf:Person">
<markbirbeck> (Because it felt less RDF-like.)
<benadida> <#foo> rdf:type foaf:Person .
<benadida> <div id="foo" rel="p" resource="o">
<markbirbeck> <a rel="license" href="http://cc....">license</a>
<inserted> Scribenick: ralphs
Mark: we're leveraging this <a href=...> example
<markbirbeck> <a id="x" rel="license" href="http://cc....">license</a>
<markbirbeck> Don't want that to change the subject of the triple.
Ben: right, adding @id to A didn't change the
interpretation
... the only time @id would matter is when neither @href or @resource appear
in the element
<benadida> <div rel="foo:bar">
<benadida> <div id="foo">
Ben: without @href or @resource, @id names the bnode subject [of triples from child elements]
<benadida> <link rel="foo:bar" href="">
<benadida> </div>
Mark: current syntax draft has mostly removed references to @id and the remaining reference is mostly wrong
Ben: @id was used before, as in the <div
id='foo'> case, to give the name '#foo' to the subject
... I may be misremembering, though
Mark: but there was a reason we removed it
Ralph: we had a long discussion of the potential confusion between the name of an HTML markup fragment and the name of some real-world resource
Ben: @rel and @rev cause chaining. @instanceof may also cause chaining.
Mark: @instanceof is special because it's choosing a different subject from <div id=>
<kwijibo> hi - there's a typo on http://www.w3.org/2006/07/SWD/RDFa/syntax/#o266637520
<kwijibo> that section keeps referring to subject resolution section 4.3 - which is the object resolution section
<RalphS> hi, kwijibo -- we're in a telecon at the moment, so a quick reply to you here might not appear
<kwijibo> np - just thought I'd mention it
<markbirbeck> http://www.w3.org/MarkUp/2004/rdf-a.html
Mark: I think it is incorrect to use @id to set the subject of child statement is wrong in all contexts
Ben: my understanding when working on http://www.w3.org/2006/07/SWD/RDFa/syntax/ was that we were moving away from @xml:id and moving to @id
<benadida> <div id="foo" instanceof="foo:Bar">
<benadida> <#foo> rdf:type foo:Bar .
Ben: is #foo the subject of a triple?
<benadida> <div about="#bar" rel="s" id="foo" instanceof="t">
<benadida> <#bar> s ? .
<benadida> ? rdf:type t .
<benadida> ? = _:div0
<benadida> ? = "#foo"
Ben: I think we agree that the object has rdf:type t.
<benadida> which one is it?
<markbirbeck> I agree that Ben has captured this correctly, and I believe that the subject of 'chained' statements should be a bnode.
<markbirbeck> <#bar> s [ rdf:type t ] .
<benadida> I think it should be:
<benadida> <#bar> s <#foo> .
<benadida> <#foo> rdf:type t .
Ralph: I never expected @id to behave the way Ben just wrote
<markbirbeck> (In the last draft that myself and Steven did, we had @nodeID to achieve this. :)
<markbirbeck> No-one liked it... ;)
<benadida> <div about="s" rel="p" resource="o">
<benadida> <span property="foo:bar">Foo!</span>
<benadida> </div>
<benadida> s p o .
<benadida> o foo:bar "Foo!" .
<markbirbeck> It's not that I saw utility in a recent email...I'm seeing that we might be able to *resurrect* this old technique.
<markbirbeck> I think the reason is that if we put link and meta inside the div, we wanted it to refer to @about.
<markbirbeck> It used to be a shorthand for:
Ben: my reasoning was that without without the
@resource, we agreed that the object of @rel='p' is a bnode and that bnode is
the subject of the child statements
... so with @resource, we just give the bnode a name
<markbirbeck> <div about="s" foo:bar="Foo!" ....
<markbirbeck> <meta property="foo:bar">Foo!</meta>
<RalphS> [I've searched http://www.w3.org/2006/07/SWD/RDFa/syntax/ for examples of @id and @rel on the same element and the only one I find also has @about which clearly seems to me to support Mark's interpretation]
Mark: some of the [shorthand] features included in old proposals have been dropped, I was wondering if we might eventually want to bring some of them back
Ben: I'm seeing more deployed examples that rely on bnode features and those should keep the interpretation if @href is added
<markbirbeck> This:
<markbirbeck> <div about="#bar" rel="s" id="foo" instanceof="t">
<markbirbeck> compared to:
<markbirbeck> <div about="#bar" rel="s" resource="foo" instanceof="t">
<markbirbeck> In other words, the use-case that we want from having @id 'name' the bnode could be achieved by using @resource.
<benadida> <div id="foo" instanceof="t">
<benadida> _:div0 rdf:type t
Ben: it seems for consistency that in this @id
@instanceof case that the subject of rdf:type is a bnode
... do we agree that @instanceof causes chaining?
Mark: it would seem that it needs to do so
<benadida> <div id="foo" instanceof="t" rel="rdf:item">
Mark: what is the minimum markup needed to
generate a triple?
... thinking about the way people understand microformats, there are special
markup triggers; e.g. class="vEvent"
... so it should not be a problem for people to also learn that @instanceof
is a trigger for interpretations of the child properties
Ben: we haven't made any new resolutions in this telecon, as we're not critical mass
ACTION: Ben writeup a summary of this discussion [recorded in http://www.w3.org/2007/07/26-rdfa-minutes.html#action01]
Ben: is this a good time for Mark and Shane to start work on the new syntax document?
Mark: yeah, I think so
[End of minutes]
Change Log:
$Log: 26-rdfa-minutes.html,v $ Revision 1.7 2007/07/26 19:50:04 swick Add visible changelog Thu Jul 26 19:42:42 2007 UTC swick Add missing regrets, mark Michael's action done