Re: Why is there no RDFa syntax working draft?

Hi Dave,

> Can you tell me why there is no working draft of the RDFa syntax?

Only because we're super swamped and have been working on solving issues
over the last few months. As Michael, Shane, and Mark have mentioned,
WDs are very much on the way, and comments have been taken from the WG
and the public for > 1 year.

We also have a number of implementations and a growing set of test cases
to make the WD complete.

> If there was a WD I'd then have somewhere to complain about:
> 1) the way plain literal, typed literals and XML Literals are used.
>    A mess.  There is no need for XML literals when simple will do.

There is actually a very clear need for XML literals in a number of
cases, but as you correctly point out there is a need to make it easy to
have plain literals, too. The latest resolution, which will be reflected
in upcoming drafts of the Primer and Syntax, is that:

<span property="dc:title">Foo</span>

yields a plain literal, while

<span property="dc:title">E=mc<sup>2</sup></span>

yields an XMLLiteral.

> 2) the lack of profile so you cannot detect RDFa (raptor as used
>    by http://triplr.org/ hits this problem)

RDFa is an extension to XHTML1.1 (and 2.0), which means you can detect
the DTD. Of course, you should probably also just try your RDFa parser,
as RDFa is meant to be copied-and-pasted into people's pages, and we've
made sure the syntax does not conflict with existing markup.

We do have a profile in the works, though it won't be mandatory: making
it mandatory would prevent the copying-and-pasting functionality we want.

> Anyway, here's Triplr reading Ivan's RDFa-FOAF into turtle
>   http://triplr.org/rdfa-turtle/http://www.ivan-herman.net/foaf.html
> Note there is some wierdness like wot:assurance [ dc:title [] a h:a ];
> I'm using a version of Fabian's XSLT to run this - I had to make
> fixes to make it use namespaces properly with libxslt.

Fabian's XSLT is not up to date. It's on my todo list to contact him.
Thanks for the gentle kick-in-the-butt ;)

> I just get the general feeling RDFa is not yet ready for
> general use and it would be risky to endorse or use now.

I'm not sure what would be so risky about endorsing it or using it now:
a number of folks are publishing RDFa, Operator 0.8 supports its
parsing, and there is a growing number of implementations. We're
definitely tight on manpower, so any help -- even if it's just testing
and beating us up when things are broken -- is much appreciated!

Your request for documents is perfectly legitimate, and we're working
hard on that. Watch the list over the next few weeks, and, if you have
time, do send us comments!

-Ben

Received on Tuesday, 4 September 2007 15:42:43 UTC