RDFA - ugly, unnecessary and offtopic (was Re: [uf-discuss] RDFa)

Chris Messina chris.messina at gmail.com
Sat May 20 06:31:13 PDT 2006

I had an interesting discussion/debate with a woman named Stephanie
who works with the w3c and had some --what I would consider -- useful
criticism of microformats -- namely in terms of general accessibility.
I believe she mentioned that RDFa would be ameliorate this problem by
providing specific tags for specific types of content.

Of course, as Tantek points out, this argument fails muster as there
are no existing UAs that support this markup (though namespacing might
be a weak counterargument to this point).

In any case, I do think that there needs to be expansion on the wiki
by folks more knowledgeable than me about implications for
accessibility that microformats offer. Additionally, I would love to
start riffing on what interfaces make sense for interacting with
microformats, both in traditional UAs and accessibility agents.

Finally, this work could lead to proposals or suggestions that
mozilla, the ie team, opera, the JAWS folks and others, could consider

Presuming that we do end up with a large number of mF implementations
across the web, what the UAs do with them is a very wide open
question. I happen to have some ideas about this and would love to
focus some SHDH or Mash Pit time on advancing our thoughts on these


On 5/19/06, Tantek Çelik <tantek at cs.stanford.edu> wrote:
> Short answer - seen it, rejected (see microformat principles), off-topic for
> microformats. (see bad topic #2 in /wiki/mailing-lists )
> Longer answer inline.
> On 5/19/06 5:31 PM, "Ben Ward" <lists at ben-ward.co.uk> wrote:
> > A W3C Working Draft published on May 16th:
> > http://www.w3.org/TR/2006/WD-xhtml-rdfa-primer-20060516/
> >
> > For Embedding RDF in XHTML.
> XHTML 2.0 to be specific.
> Which is not "XHTML" by any common use of the term on the Web.
> Typical use of XHTML implies XHTML 1.0, nothing more.
> > Gives iCal and vCard examples. In
> > practice, there's a bit mark-up involved than with µF (namespace
> > declarations for a start) but seems to acheive much the same thing in
> > the end.
> Uglier and achieves nothing more.
> > One interesting thing to note is the use of the META element for
> > embedding computer-readible data, e.g.
> >
> > <meta property="cal:dtstart" content="20060508T1000-0500">May 8th at
> > 10am</meta>
> Interesting is one way to put it.  It's also:
> * Invalid in today's XHTML.
> * Not CSS friendly (hint: classes are much easier to use with style rules)
> * Makes the same mistake as XML ('property' attribute can only take one
> value, unlike 'class' which is a multivalued set.  They will encounter this
> problem with their examples as soon as they try to markup a vCard for
> someone with a middlename for example.
> * Misleading/unnecessary use of <meta> - there is nothing "meta" about the
> datetime of an event, that is information core to the event.
> > Has this ever been considered for Microformats?
> Yes, I've been familiar with and tracking RDF/A ever since it was first
> proposed.
> You might even say that the insanity of the ugliness of RDF/A was one of
> inspirations for a lot of simplicity and human-centric aspects of
> microformats.
> > It seems especially
> > relevant given the recent uncertainty regarding ABBR/@TITLE and
> > accessibility tools.
> No uncertainty there, that's a resolved issue, see URL in previous message
> on the subject.  Anything else is just FUD.
> > I have to admit, I've never seen META used
> > outside the HEAD and never even considered it as valid.
> That's because it's not valid, neither in HTML4.x nor in XHTML1.x.
> > Could it be a
> > viable alternative (if it's considered that an alternative is ever
> > needed).
> I don't think it is viable no.  I put RDF/A in the bucket of interesting
> science projects doomed to failure in the real world, and certainly not even
> worth the time to discuss on this list.
> Thanks,
> Tantek
> P.S. For extra nitpickers:
> * The use of QNames is *NOT* a use of "standard XML namespaces", not by a
> long shot. QNames don't work with CSS Selectors, thus being impractical for
> presentation, thus failing to satisfy the primary use of semantic markup.
> * The fact that this draft had to invent a new form of URI (CURIE) should be
> a strong indicator that there is something wrong.  Whenever you find
> yourself inventing new piece of technology for an orthogonal part of the
> stack, it usually means you're doing something wrong in your layer.
> _______________________________________________
> microformats-discuss mailing list
> microformats-discuss at microformats.org
> http://microformats.org/mailman/listinfo/microformats-discuss

More information about the microformats-discuss mailing list