Re: RDFa + RDF/XML Considered Harmful? (was RE: Ordnance Survey data as Linked Data)

An important difference between embedded RDF/XML and RDFa is that RDFa and the xhtml can use the same "literals".  In other words, the text viewed by the human, and the text stored as the literal object of a triple is the same.  xhtml has the <meta/> tags which are pretty much ignored because they often have nothing to do with the real content of the page.

If I have a web page of for sale listings, the RDFa isn't meta-data about the listings, it is the data.  It's just data-data, the data we really want, not extra data we thought to hide from human eyes.  And that has implications for how useful, trustable, and spamable the content may be.

And in the cases where the RDFa adds semantics on top of what the human sees, it's position carefully in place with the xhtml it modifies, not at the top or bottom of the document, relagated to 2nd class "meta-data" status.

Taylor

----- Original Message ----
From: Mark Birbeck <mark.birbeck@webbackplane.com>
To: Bijan Parsia <bparsia@cs.man.ac.uk>
Cc: public-lod@w3.org; SW-forum Web <semantic-web@w3.org>
Sent: Tuesday, July 15, 2008 7:05:52 AM
Subject: Re: RDFa + RDF/XML Considered Harmful? (was RE: Ordnance Survey data as Linked Data)


Hi Bijan,

> One funny aspect of RDF/XML, as I understand the history, is that some of
> the quirkier aspects of its design stemmed from the goal of being embedable
> in HTML (hence all the alternative forms) in a legacy browser compatible
> way.

That's interesting, I'd not heard that.

I did think though, that one of the things about the RDF/XML structure
was an attempt to enable many XML layouts to be interpreted as RDF.
But obviously that's enormously difficult.


> In this sense, RDFa aims to fulfill one of the RDF/XML goals that helped
> make it (somewhat) unsuitable to backend work (compare with NTriples).
>
> If at first you don't succeed, try again 10 years later ;)

Work actually began on RDFa 5 years ago, so the wait wasn't quite so
long before 'trying again'. :)

But it's interesting that you say all of this, since my first attempt
at putting metadata into HTML was to actually employ a subset of
RDF/XML, applied to HTML documents.

Part of the parsing rules for RDF/XML say that parsing can begin when
an @rdf:about is discovered, so I tried things like this:

  <html>
    <head>
      ...
    </head>
    <body>
      <div rdf:about="blah">
        ...
      </div>
    </body>
  </html>

The elements prior to the @rdf:about are ignored in RDF/XML parsers,
so it seemed that it might be possible to create XHTML+RDF/XML hybrids
this way. But ultimately, getting predicates was way too awkward, and
since the primary goal was to make it easy for people to publish
metadata, this was a show-stopper.

(Not to mention that having to use namespace prefixes on the attribute
names would probably have killed the whole thing.)

In the end it became easier to create a new set of HTML-specific
attributes that played the same role as the RDF/XML ones, but over
which we had complete control in terms of their meaning, position,
etc.

Regards,

Mark

-- 
Mark Birbeck, webBackplane

mark.birbeck@webBackplane.com

http://webBackplane.com/mark-birbeck

webBackplane is a trading name of Backplane Ltd. (company number
05972288, registered office: 2nd Floor, 69/85 Tabernacle Street,
London, EC2A 4RR)

Received on Tuesday, 15 July 2008 12:40:48 UTC