Re: is this valid to make a named graph in RDFa?

HI Ivan,

Just a small point on this:

>  There are some features that RDFa has not defined, ie, not defined to be
>  automatically generated, and reification is one of those. Simply put,
>  there was no real demand for it...

We did used to have reification (and that probably explains why Golda
later in this thread says 'I'm sure someone told me that reification
was possible in RDFa').

It was marked up like this:

  <link rel="p" href="o">
    <link rel="whosaidit" href="#me" />
  </link>

  <meta property="p" content="o">
    <link rel="whosaidit" href="#me" />
  </meta>

In short, link and meta placed as children of link and meta, created
reified statements. (It's only when applied to link and meta though; a
link or meta as a child of statements on a span or div worked in just
the same way that a span or div would.)

But it turned out that all current browsers do odd things with link
and meta when placed in the body of the document, so we were not able
to support this syntax for reification in XHTML+RDFa (at least not if
we wanted it to work in current browsers). And since the RDF community
as a whole doesn't seem to be entirely convinced that reification is
the right way to go, we decided that there wasn't exactly pressure to
try to find an alternative reification syntax, and so the whole thing
was eventually dropped.

Speaking for myself, I'm not that bothered. :) I have never liked
reification, and I'm a big fan of the named graph approach. The
problem with RDF/XML is that you could never 'name the graph' because
an RDF/XML document is nothing but metadata...in part why we have the
whole 'information resource' v. 'resource' question. But RDFa is
slightly different.

An XHTML+RDFa document is both a document (an information
resource)_and_ some metadata (some resources), and for that reason you
could regard the URL of the document as the name of the graph.

And using the nice separation that HTML has of head and body, we could
put statements about the graph/document in the head (who created it,
when, and so on), and the statements themselves in the body:

  <html>
    <head about="">
      <title>...</title>
      <meta ...stuff about the graph, like who created it .../>
    </head>
    <body about="#">
      <span ...stuff about a person, like where they work .../>
    </body>
  </html>

You could even add an rdf:type for named graph, which I think Jeremy
proposed elsewhere:

  <html>
    <head about="" instanceof="rdf:Graph???">
      <title>...</title>
      <meta ...stuff about the graph, like who created it .../>
    </head>
    <body about="#">
      <span ...stuff about a person, like where they work .../>
    </body>
  </html>

or just imply this automatically when generating triples.

Anyway, I don't claim to have worked out all of the ins and outs of
this...but it certainly seems quite compelling, and it I think it
neatly resolves the information resource questions that often get
raised.

Regards,

Mark

-- 
  Mark Birbeck

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

  x-port.net Ltd. is registered in England and Wales, number 03730711
  The registered office is at:

    2nd Floor
    Titchfield House
    69-85 Tabernacle Street
    London
    EC2A 4RR

Received on Thursday, 6 March 2008 20:46:35 UTC