RE: RDFa and its relationship to XHTML

On 08.06.2006 19:07:27, Mark Birbeck wrote:
>I think it's quite natural to extract:
>
> <http://www.w3.org/>
>  <http://www.w3.org/Copyright>
>  <http://www.w3.org/Consortium/Legal/ipr-notice#Copyright>
> .
>
Hmm, then I strongly suggest to change the RDFa spec. It
should not be possible to generate triples from unqualified
rel/rev values. There is simply too much HTML out there that
would lead to incorrect or at least unwanted statements.

RDF in HTML is currently discussed almost every day, the demand
is clearly there. However, the more I look at RDFa, the more I
think it's heading in the wrong direction. It should be compatible
with existing HTML documents at least. From reading comments and
impressions on the Web, some other suggestions would be to
- drop the whole CURIE stuff,
- change the IRI abbreviation mechanism. The current encoding
  is identical to qnames which shouldn't appear in attribute. This
  does not only confuse XML developers. Look at eRDF, it uses a
  "prefix-localname" notation which makes is still similar enough
  to qnames, disambiguating enough to not produce unwanted triples,
  and can also nicely be used in a stylesheet. Am RDFa-specific
  notation would also give free syntax options, e.g. you could
  declare "taxo-1234" as valid and solve the CURIE use cases.
- change the namespace declaration process. Again, look at eRDF which
  uses the already deployed '<link rel="schema" href="..." />'
  practise. The xmlns mechanism isn't supposed to work for attribute
  values anyway, and it also does not and will not work for XHTML1.x
  unless you change the specifications.

Or, alternatively, just don't claim RDFa to be XHTML1-happy.

>RDFa then builds on this to say, if we like that 'interpretation' of @rel
>and @rev (and it's not very controversial), why not give the 'predicate'
>part a bit more power:
>
>  <a
>   rel="cc:license"
>   href="..."
>  >Attribute, sharealike, no-moustaches-drawn-on-faces</a>
>
But this example requires that cc is declared somewhere, and
the current way this is done in RDFa breaks XHTML1 conformance.
This example doesn't work.

>But whatever you think of the last two 'steps', the first step is pretty
>much given by HTML 4.01, and that is why I'm flagging up a problem with the
>way that @rel has been used in other embedded formats.
Yeah, it is a problem for RDFa, and as it's impossible to change the
current usage of @rel, we better change the RDFa spec, no?

Ben

>
>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 18:59:42 UTC