Re: Non-XHTML host languages for RDFa

On Tue, 2009-12-01 at 23:57 +0100, Christoph LANGE wrote:
> 2009-12-01 18:17 Toby Inkster <tai@g5n.co.uk>:
> > 1. Allows lists of keywords to be configured on a per-attribute basis
> > for @rel, @rev, @property, @datatype, @typeof, @about, @resource and
> > @graph.
> 
> aha, that sounds like an interesting approach, to enable that per
> attribute. Not sure whether it should really be different for, e.g.
> @rel and @rev, and whether one would want keywords on @about and
> @resource, but, well, why not? Let's see how people use it.

I imagine that people would define @rel and @rev the same. As it's a
structure that uses pointers (a hashref in Perl), it's pretty easy to
set @rev's keywords list to be a pointer to @rel's. But they can be used
separately if need be.

> (What is @graph BTW?)

RDF::RDFa::Parser has experimental support for quad-based RDFa parsing
as per <http://buzzword.org.uk/2009/rdfa4/spec>. (Support is disabled by
default.)

> > 4. Supports colon-less CURIE prefixes (e.g.
> > xmlns:foafname="http://xmlns.com/foaf/0.1/name"
> property="foafname").
> 
> I have always thought that a name_space_ is something that contains a
> lot of localnames.  So this is indeed a creative (ab)use of
> "namespaces" that only contain a single entity.  I wouldn't say it is
> illegal, though, it's definitely an interesting and elegant hack. 

It's legal in RDFa to do this:

xmlns:foafname="http://xmlns.com/foaf/0.1/name" property="foafname:"

Enabling the colon-less option just allows that last colon to be
omitted. This idea comes from Mark Birbeck's article here:
<http://webbackplane.com/mark-birbeck/blog/2009/04/30/tokenising-the-semantic-web>

-- 
Toby A Inkster
<mailto:mail@tobyinkster.co.uk>
<http://tobyinkster.co.uk>

Received on Wednesday, 2 December 2009 09:21:38 UTC