W3C home > Mailing lists > Public > public-rdf-in-xhtml-tf@w3.org > December 2007

Re: [PROPOSAL] Named graphs in RDFa

From: Mark Birbeck <mark.birbeck@x-port.net>
Date: Wed, 19 Dec 2007 20:53:22 +0000
Message-ID: <a707f8300712191253m51904011tf5cd5070f2a5d7f3@mail.gmail.com>
To: "Hausenblas, Michael" <michael.hausenblas@joanneum.at>
Cc: public-rdf-in-xhtml-tf@w3.org, "SWD WG" <public-swd-wg@w3.org>

Hi Michael,

> Thanks for your explanation - as always a pleasure to read.
> If I understood correctly this is basically what I was looking for! :)
>
> I therefore withdraw my original proposal and would like to suggest
> that this aspect (or more precisely: your explanation of it ;)
> somehow finds its way into the RDFa Syntax document.

This is still in the realm of theory, you understand. :)

But if people wanted to go this route, then all that would be needed is:

  * to indicate the the 'default value' for @about in <body> was
slightly different
    to the 'default value' for @about that we currently have (i.e, the
current document
    URI);

  * that when triples are produced, we stipulate that the 'name' of
the graph is the
    URI for the document;

  * that the foaf:primaryTopic 'defaults' to the resource section.

This would give us two distinct lots of triples, one for URI "xyz" and
one for URI "xyz#". If the RDF world decides that a URI can have an
rdf:type of rdf:graph, then we'd just add that triple, and all would
be well.

To illustrate, this:

  <html>
    <head>
      <meta property="dc:creator">Mark Birbeck</meta>
    </head>
    <body>
      <span property="foaf:name">Mark Birbeck</span>
    </body>
  </html>

would yield:

  <> rdf:type rdf:graph .
  <> dc:creator "Mark Birbeck" .
  <> foaf:primaryTopic <#> .

  <#> rdf:type foaf:Person .
  <#> foaf:name "Mark Birbeck" .

and so on.

That may not be the total answer, and people may feel that too much
defaulting is going on. But my feeling is that there is a solution in
here somewhere, fighting to get out, and the head/body distinction is
probably key to it.

(As it happens, even if we did none of this in RDFa, if RDF was to
gain an 'rdf:graph' type, then we could make use of it in the way I've
described, but we'd simply have to spell it out.)

Regards,

Mark

On 19/12/2007, Hausenblas, Michael <michael.hausenblas@joanneum.at> wrote:
>
> Mark,
>
>
> Cheers,
>         Michael
>
> ----------------------------------------------------------
>  Michael Hausenblas, MSc.
>  Institute of Information Systems & Information Management
>  JOANNEUM RESEARCH Forschungsgesellschaft mbH
>  Steyrergasse 17, A-8010 Graz, AUSTRIA
> ----------------------------------------------------------
>
> > -----Original Message-----
> > From: Mark Birbeck [mailto:mark.birbeck@x-port.net]
> > Sent: Wednesday, December 19, 2007 5:01 PM
> > To: Hausenblas, Michael
> > Cc: public-rdf-in-xhtml-tf@w3.org; SWD WG
> > Subject: Re: [PROPOSAL] Named graphs in RDFa
> >
> > Hi Michael,
> >
> > This is an interesting problem, and one I've avoided
> > commenting on because it could take every waking minute of
> > the day. :) Anyway, all of the following is in the
> > 'theoretical' bin....
> >
> > I think RDFa is unusual in that it doesn't have a named graph
> > problem in the way that, say, RDF/XML does, because the XHTML
> > document *is* the 'named graph'.
> >
> > The interesting thing about an HTML/XHTML document is that
> > there is both a head and a body section of the document,
> > which means that we could define things in such a way that we
> > are making *two* lots of statements; one lot about the
> > document itself, and one lot about the 'content' referred to
> > by the document.
> >
> > The document itself is a named graph, since it has a URL, and
> > contains metadata. Of course, normally all metadata in the
> > document is 'about'
> > that document unless overridden by an @about, so to
> > distinguish between statements about the document (i.e., the
> > graph) and statements about some other things you'd always
> > have to add an @about.
> >
> > However, if a typical document looked like this:
> >
> >   <html>
> >     <head about="">
> >       ...statements about the named graph...
> >     </head>
> >     <body about="#">
> >       ...statements about the content, e.g., my FOAF page...
> >     </body>
> >   </html>
> >
> > you would effectively have a 'named graph', which is the
> > information resource at "", accompanied by a resource,
> > identified by "#"; note how this might also solve the
> > interminable information resource question.
> >
> > (As it happens this trick of distinguishing between the
> > document carrying the metadata and what the metadata is about
> > could have been done with RDF/XML, but in RDF/XML if
> > @rdf:about is empty it means that statements are being made
> > about the current document; in short, there is no way to
> > distinguish between the document that contains the graph, and
> > the graph itself.
> >
> > If we also added a statement that the 'primary topic' of the
> > named graph, was the resource identified in the body:
> >
> >   <html>
> >     <head about="">
> >       <link rel="foaf:primaryTopic" href="#" />
> >       ...statements about the named graph...
> >     </head>
> >     <body about="#">
> >       ...statements about the content, e.g., my FOAF page...
> >     </body>
> >   </html>
> >
> > then everything is complete; now, when someone links to the 'graph'
> > from an HTML page (the most likely scenario), it is a simple
> > matter to sort out what exactly is being referred to.
> >
> > Regards,
> >
> > Mark
> >
> > On 19/12/2007, Hausenblas, Michael
> > <michael.hausenblas@joanneum.at> wrote:
> > >
> > >
> > > All,
> > >
> > > Proposal:
> > > Add a mechanism to RDFa that allows to indicate a graph's
> > provenance
> > > (aka named graphs, etc.).
> > >
> > > In the first phase we should decide IF we want this (or not
> > :), then
> > > if/when we agree, we should discuss HOW to implement it.
> > >
> > > Rational:
> > > Due to the ongoing discussions [1] and Fabien's W3C member
> > submission
> > > [2] is ask myself:
> > > Why don't we introduce this feature in RDFa?
> > >
> > > As we are about (or partly already did) add new stuff
> > anyway (such as
> > > XHTML voc, CURIE, etc.), why not supporting this IMO very important
> > > feature right from the beginning?
> > >
> > > Any thoughts?
> > >
> > > Cheers,
> > >         Michael
> > >
> > > [1]
> > http://lists.w3.org/Archives/Public/semantic-web/2007Dec/0035.html
> > > [2] http://www.w3.org/Submission/rdfsource/
> > >
> > > ----------------------------------------------------------
> > >  Michael Hausenblas, MSc.
> > >  Institute of Information Systems & Information Management
> > JOANNEUM
> > > RESEARCH Forschungsgesellschaft mbH  Steyrergasse 17, A-8010 Graz,
> > > AUSTRIA
> > >
> > >  <office>
> > >     phone: +43-316-876-1193 (fax:-1191)
> > >    e-mail: michael.hausenblas@joanneum.at
> > >       web: http://www.joanneum.at/iis/
> > >
> > >  <private>
> > >    mobile: +43-660-7621761
> > >       web: http://www.sw-app.org/
> > > ----------------------------------------------------------
> > >
> > >
> >
> >
> > --
> >   Mark Birbeck, formsPlayer
> >
> >   mark.birbeck@formsPlayer.com | +44 (0) 20 7689 9232
> >   http://www.formsPlayer.com | http://internet-apps.blogspot.com
> >
> >   standards. innovation.
> >
>


-- 
  Mark Birbeck, formsPlayer

  mark.birbeck@formsPlayer.com | +44 (0) 20 7689 9232
  http://www.formsPlayer.com | http://internet-apps.blogspot.com

  standards. innovation.
Received on Wednesday, 19 December 2007 20:53:43 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Tuesday, 8 January 2008 14:15:19 GMT