Re: POWDER-related questions about RDFa linkage

Phil Archer wrote:
> 
> Thanks Ivan, most helpful.
> 
> I'll add the RDFa URI profile to the example.
> 
> My example 2, with link rel="powder" is not related to RDFa - that's
> just straight HTML+POWDER, the point being that, if all we're doing is
> linking a doc to a Description Resource, RDFa's potential is probably
> being under-used.
> 
> Your example of how to use rev and about on a hyperlink is just what I
> was hoping for, thanks. If I add that in the Rec Track doc as another
> example, will any of this affect our anticipated timeline of LC later
> this month and CR before year's end?
> 

Well.. the goal of the RDFa TF is to put out a last call before the end
of the year. That means your dependency (ie, RDFa) will be one step
behind which is o.k. per process. By doing so, you just increase the
fire under our own noble side to keep to the RDFa schedule, so that
should be o.k.:-)

Ivan



> Phil.
> 
> (P.S. When I say grandly that I'll add this and that to the doc, what I
> mean is, I'll add it to the draft so it can be discussed by the WG next
> week!)
> 
> Ivan Herman wrote:
>> If you allow me to impersonate Michael for the moment:-) (which does not
>> mean that he cannot answer!:-)
>>
>> Phil Archer wrote:
>>> Michael,
>>>
>>> As you and others have been kind enough to look at the potential usage
>>> of RDFa and POWDER, I'd like to ask a couple of questions please.
>>>
>>> I'd like to significantly improve what we say about RDFa, either in the
>>> Rec Track Description Resources doc [1] or, if more appropriate, in a
>>> non-normative POWDER Primer we'll be starting work on later this month.
>>>
>>> Easier question first. Can I ask you for a sanity check on this example
>>> please:
>>>
>>> Example 1
>>> =========
>>>
>>> <html
>>>   xmlns="http://www.w3.org/1999/xhtml"
>>>   xmlns:wdr="http://www.w3.org/2007/05/powder#"
>>>   <head>
>>>     <title>The English Civil War</title>
>>>     <link rel="wdr:describedBy"
>>> href="http://education.example.org/powder.rdf#DR_1" />
>>>   </head>
>>>   <body>
>>>     …
>>>     <p>Charles I came to the throne believing in his Divine Right to
>>> rule...
>>>     …
>>>   </body>
>>> </html>
>>
>> Although we do not have it in the current document yet, we will have an
>> RDFa profile, too. It may not be mandatory to use it, but advised. Ie,
>> you should probably add the
>>
>> <head profile="http://www.w3.org/ns/rdfa/">
>>   ....
>>
>> Otherwise, this should be o.k.; it will generate the triple
>>
>> <URI-of-the-document> wdr:describedBy
>> http://education.example.org/powder.rdf#DR_1.
>>
>>> That is, a single use of wdr:describedBy at document level. Actually,
>>> for this kind of linkage from an HTML doc, the usual way will be to use
>>> an HTML profile we're drafting and a relationship type of powder thus:
>>>
>>> Example 2
>>> =========
>>>
>>> <html xmlns="http://www.w3.org/1999/xhtml">
>>>    <head profile="http://www.w3.org/2007/10/powder-profile">
>>>       <link rel="powder"
>>>          href="http://education.example.org/powder.rdf#DR_1" />
>>>       <title>The English Civil War</title>
>>>    </head>
>>>    <body>
>>>       <p>Charles I came to the throne believing in his Divine Right to
>>> rule...</p>
>>>    </body>
>>> </html>
>>>
>>
>> Actually... this is still in flux. At the moment, the RDFa rule is that
>> @rel values without a prefix will be ignored. We do not have (yet?) a
>> general mechanism that would allow an RDFa processor to understand a
>> profile in the head automatically and take those @rel values. Ie, it may
>> be better to stick with the wdr: prefix. (As I said: this is still a bit
>> in flux, the TF members may well chime in here!)
>>
>>> But I'm really hoping that RDFa can be used for more powerful things. In
>>> particular, that we can use it to link to a Description Resource that
>>> describes the target of a hyperlink as simply as possible. If I
>>> understand RDFa correctly that means doing something like this:
>>>
>>> Example 3
>>> =========
>>>
>>> <html
>>>   xmlns="http://www.w3.org/1999/xhtml"
>>>   xmlns:wdr="http://www.w3.org/2007/05/powder#"
>>>   <head>
>>>     <title>The English Civil War</title>
>>>     <link rel="wdr:describedBy"
>>> href="http://education.example.org/powder.rdf#DR_1" />
>>>   </head>
>>>   <body>
>>>     …
>>>     <p>Charles I came to the throne believing in his
>>>     <link about="#divRight" rel="wdr:describedBy"
>>>          href="http://education.example.org/powder.rdf#DR_2">
>>>     <a href="http://education.example.org/divine_right.html"
>>>          id="divRight">Divine Right</a> to rule...
>>>     …
>>>   </body>
>>> </html>
>>>
>>
>> First of all: RDFa does not change the content model of XHTML1, ie, the
>> <link> element is _not_ permitted in the body.
>>
>>
>> <html
>>    xmlns="http://www.w3.org/1999/xhtml"
>>    xmlns:wdr="http://www.w3.org/2007/05/powder#"
>>    <head profile="http://www.w3.org/ns/rdfa/">
>>      <title>The English Civil War</title>
>>      <link rel="wdr:describedBy"
>> href="http://education.example.org/powder.rdf#DR_1" />
>>    </head>
>>    <body>
>>      …
>>      <p>Charls I came to the throne believing in his
>>      <a
>>        about="http://education.example.org/powder.rdf#DR_2"
>>        rev="wdr:describedBy"
>>        href="http://education.example.org/divine_right.html">Divine
>> Right</a> to rule...
>>
>>      …
>>    </body>
>> </html>
>>
>> The trick is the usage of @rev and not @rel; @rev, essentially reverts
>> the roles for subject and object. I guess this is what you want, right?
>>
>> I hope this helps.
>>
>> Ivan
>>
>>
>>
>>
>>> So we have a hyperlink in the document to one about divine right. That
>>> second document is the subject of a triple thus:
>>>
>>> <http://education.example.org/divine_right.html>
>>>   wdr:describedBy
>>> <http://education.example.org/powder.rdf#DR_2>
>>>
>>> is there a more terse way of creating this triple within the RDFa/XHTML?
>>>
>>> I guess the problem is that would entail having two href attributes in a
>>> single hyperlink, which is silly, but the use cases are pretty clear for
>>> us. For example, imagine that the English Civil War document is mobileOK
>>> and is being displayed on a mobile device - should my browser display
>>> the hyperlink to the divine right document or not? If it's mobileOK too
>>> (as declared in DR_2) then the answer's yes, if not, then just show the
>>> text without the hyperlink.
>>>
>>> Finally, how far can we go in a Rec Track document that we're hoping to
>>> declare as at Last Call after next week's face to face? Is Example 1
>>> sound? Is Example 3 sound?
>>>
>>> Thanks very much for your help,
>>>
>>> Phil.
>>>
>>> [1] http://www.w3.org/TR/2007/WD-powder-dr-20070925/#semlink
>>>
>>
> 

-- 

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

Received on Thursday, 1 November 2007 12:35:31 UTC