W3C home > Mailing lists > Public > public-rdf-wg@w3.org > May 2012

Re: rdf-spaces

From: Andy Seaborne <andy.seaborne@epimorphics.com>
Date: Thu, 10 May 2012 14:26:14 +0100
Message-ID: <4FABC1F6.1060502@epimorphics.com>
To: public-rdf-wg@w3.org


On 09/05/12 15:59, David Wood wrote:


>> It's the named graphs part that I'm missing. (JSON does not seem to
>> help so I guess I didn't understand your initial remark properly.)
>
> Sure.  The Turtle-in-HTML section provides a way to represent triples
> in HTML.  But where did those triples come from?  How can I refer
> back to the graph?

I may be jumping to a conclusion but if there are triples in a doc at
<URL> it's just like pure TTL at <URL>.   You refer back to the place 
where they were.  This assumes that multiple such blocks are the merge.

Otherwise, I'd be included to use TriG, c.f. TTL vs TriG.

> For example, I may wish to display those triples on a Web page and
> provide a link to the graph for provenance.  Alternatively (my more
> common use case), I may wish to present the triples in an HTML form
> and cause a graph update upon submission of the form.  In that case,
> I must know the graph identifier in order to perform the update.

Agreed.  I thin it's another step to have the graph identifier actually 
in the serialization ... just like a TTL file on the web does not 
contain it's URL (and of course one doc can have several URLs, and in 
fact always does because of equivalent ways of writing the same URL .. 
for some definition of "same".  /me Hole - stop digging.).

> We currently do that with RDFa and a server-side template engine, but
> we can't use that mechanism in all cases, e.g. when the requested
> resource wishes to use RDF that can't be directly navigated from the
> resource - a data island.  So, we end up defining "named" queries
> (SPARQL queries that have URLs and, when poked, emit JSON via a
> transform).  It would be great to short-circuit that rather
> inefficient process (and common use case, at least for us) by
> allowing relevant datasets to be immediately accessible in the HTML
> for processing via Javascript.

Slight worry that we can going from "Turtle in HTML, a suggestion" to an 
almost REC-grade "Turtle in HTML like RDFa is triples in HTML".  Next 
will be the Turtle referring to concepts in the document it is in.

Fine thing to have ... time consuming ... later.

	Andy

>
> Regards, Dave
>
>
>>
>> Andy
Received on Thursday, 10 May 2012 13:26:49 GMT

This archive was generated by hypermail 2.3.1 : Tuesday, 26 March 2013 16:25:48 GMT