Re: HTML 4 Profile for RDFa

...snipses
>
> Given that people will see a spec for XHTML and emulate that in 
> faus-XHTML served as text/html (e.g., Drupal), I would hope that one 
> of the following hold true:
>
> 1) It obviously and clearly doesn't work: either the syntaxes are 
> wildly divergent, or no triples are ever produced.
>
> 2) There exists a set of best practices which, if followed, means that 
> conforming examples will work in either context.  Examples: always use 
> 'xmlns' (even if XMLns might work in HTML), or always use lowercase 
> prefixes (even if upper or mixed case prefixes may work in XHTML).
>
> At the moment, it appears that option 1 is out of the question.  I 
> would submit that option 2, while it doesn't /technically/ change the 
> definition of XML-based languages like RDFa for XHTML, it does 
> retroactively impose a set of best practices on such usage.
>

+1

An issue that hasn't been brought up, though, is that there is an 
assumption going forward that attributes RDFa is dependent on don't get 
formally redefined in HTML5 or a future version of HTML. @property came 
very close to being redefined in HTML5, which would have had a seriously 
negative impact on the future of RDFa in X/HTML5. As it is, the section 
on generating RDF from Microdata could have serious impact on larger 
pools of RDF data, by introducing variances in the underlying data model.

In this discussion, we've been focusing purely on RDFa, as it works with 
HTML today. This is good, if for no other reason that it could lead to 
an excellent Best Practices guide. I believe, though, that it would be 
fair and equitable for the discussion to encompass what must occur to 
ensure that RDFa works with the HTML of tomorrow.

> - Sam Ruby
>
>
Shelley

Received on Sunday, 24 May 2009 13:20:50 UTC