RE: Finding RDFa content on the web

Hi Tim,

> My question is, are there Google queries that will be useful 
> for finding XHTML documents with RDFa content?  For example, 
> a Google query file 'rdf -rss filetype:rdf' produces lots of 
> RDF documents. I tried searches like '"rel=" "html xmlns:"'
> but virtually all of the of the documents found are using 
> conventional uses of the rel attribute.

I don't think we will find such a query, unfortunately. I'll explain why,
and also use this as an opportunity to comment on some of the
misunderstandings in discussions about RDFa, microformats, etc., about just
how 'different' RDFa is.

The use of the @rel and @rev attributes is already standard HTML. All we
have done in RDFa is pinned down what it means in terms of RDF triples. The
problem with the very specific approaches that microformats and GRDDL take
is that you can't know what triples are generated by @rel/@rev values,
without looking at the processor or the syntax. This in effect breaks the
standard HTML usage, since the attributes qualify the links and are
therefore something about the document's relationship to the external
resource.

That's not to say we haven't extended it--we've said that the value in
@rel/@rev can be a CURIE (although that doesn't actually violate HTML
syntax) and we've also said that the addition of the @about attribute allows
you to use @rel and @rev as predicates of some subject other than the
current document.

But since RDFa is an extension of 'best practice' HTML authoring, then it
will be difficult to find a search pattern that finds documents that are
'consciously' using RDFa, rather than finding documents that are just 'good'
XHTML.


> If anyone has suggestions for search engine queries that 
> might be good at finding RDFa content, please let me know.
> If there aren't any, maybe it would be good to develop a 
> convention by which an XHTML document can assert that it has 
> RDFa content and to encourage it's use as a best practice.

I sympathise, but to me it sort of goes against the grain of the gentle
ramp-up that you can get from RDFa, as you build on normal HTML authoring.
The 'levels' go like this:

  1. Use @rel and @rev to give information about your links.

So far, just standard HTML. (Best practice would say that a @profile is
used, which RDFa needs to say something about.)

  2. Use namespace-prefixed @rel and @rev values to make
     this information globally understandable.

Using namespaces would still be valid XHTML, as it happens, although the
concept may be 'new' for some authors.

  3. Use @property to provide predicates for inline mark-up.

This is new, but not that difficult--it's just the equivalent of @rel and
@rev for inline text.

  4. Use @about to indicate that something other than the
     document is the subject.

Also new, but intuitive--it sets a context for a block of metadata that
concerns some external resource, such as an image, a sound file, a license,
a friend, etc.

  5. Use <meta> and <link> anywhere in the document.

Also new...also intuitive...create metadata about items in your document
(like quotes, images, paragraphs of text).

As you can see, RDFa tries to gently build on HTML best practice, but ramps
up to full-blown RDF (we've sneaked reification and bnodes in there!). But
at what point you say you are using RDFa, I don't know.

Regards,

Mark


Mark Birbeck
CEO
x-port.net Ltd.

e: Mark.Birbeck@x-port.net
t: +44 (0) 20 7689 9232
b: http://internet-apps.blogspot.com/
w: http://www.formsPlayer.com/

Download our XForms processor from
http://www.formsPlayer.com/

Received on Thursday, 8 June 2006 10:42:50 UTC