[Fwd: Re: Request to publish HTML+RDFa (draft 3) as FPWD]

(cc'ing the HTML WG mailing list)

Forwarded message 1

  • From: Peter Mika <pmika@yahoo-inc.com>
  • Date: Tue, 22 Sep 2009 21:59:14 +0200
  • Subject: Re: Request to publish HTML+RDFa (draft 3) as FPWD
  • To: public-rdf-in-xhtml-tf@w3.org
  • CC: Kevin Haas <khaas@yahoo-inc.com>, Evan Goer <goer@yahoo-inc.com>
  • Message-ID: <4AB92C92.8010303@yahoo-inc.com>
  • Archived-At: <http://www.w3.org/mid/4AB92C92.8010303@yahoo-inc.com>
Dear All,

I'm not intending to jump into the subject matter of this discussion, 
just wanted to set the record straight with respect to the email below 
and the referenced email of Ian [1].

The RDFa implementation used in our indexer is to the best of our 
knowledge complete and correct with respect to the RDFa specification 
with the exception of predicates in the xhtml namespace (which we 
ignore). This is what feeds SearchMonkey, the new Yahoo! search page 
with object drill-down functionality and the BOSS API.

The SearchMonkey object validators referenced by Ian are not RDFa 
complete in that there are various ways to write down the same set of 
triples that will not pass these validators. However, these validators 
are intended for webmasters who are new to RDFa and are looking for at 
least one correct way to embed RDFa markup into their pages. These 
validators give them the signal that they have produced correct RDFa 
(even if they could have done it differently). Put it differently, these 
are not RDFa validators and they don't claim to be either.

Of course, if anyone finds any bugs, please let us know. We equally 
welcome any other direct feedback, it doesn't have to be negative ;)



-------- Original Message --------
Subject: 	Re: Request to publish HTML+RDFa (draft 3) as FPWD
Resent-Date: 	Tue, 22 Sep 2009 15:55:48 +0000
Resent-From: 	public-rdf-in-xhtml-tf@w3.org
Date: 	Tue, 22 Sep 2009 18:55:00 +0300
From: 	Henri Sivonen <hsivonen@iki.fi>
To: 	Mark Birbeck <mark.birbeck@webbackplane.com>
CC: 	HTMLWG WG <public-html@w3.org>, RDFa mailing list 
References: 	<4AB06B49.9020905@digitalbazaar.com> 
<02670203-4E4F-44E3-8848-09436F5D313B@iki.fi> <4AB87D39.1040102@gmx.de> 

On Sep 22, 2009, at 17:37, Mark Birbeck wrote:

> On Tue, Sep 22, 2009 at 12:17 PM, Henri Sivonen <hsivonen@iki.fi>  
> wrote:
>>> The BBC is publishing RDFa in the form of program reviews.
> You missed one.
> Nothing to say on this?

I quoted it above the UK government bit to make the same comment about  

>>> UK government websites are publishing job vacancies and  
>>> consultations with
>>> RDFa.
>> Who consumes this data? (My point being: If a cow falls in the  
>> forest but no
>> one is there to observe it, does it make a path?)
> Unless it flew there...yes, it does make a path.
> But seriously...
> ...oh, why bother.
> If the UK government publishing a ton of metadata in the form of RDFa
> still only puts us at roughly the same level of adoption as Microdata
> I think we've reached the end of rational debate.

Until it's also shown that the data is being consumed, it hasn't been  
shown that the publication of data resulted in communication. I'm not  
saying that no one is consuming it. I'm asking if anyone is. I think  
it's an important point in assessing the success of a Web technology  
if it actually enables communication. Producers without consumers  
aren't interesting in their own right.

>>> Drupal 7 includes RDFa support.
>> What does that mean? Does Drupal output RDFa? If so, who consumes  
>> it? Does
>> it ingest RDFa? If so, from where?
> Who cares?

I care, obviously, since I asked.

Those are very relevant questions. We've seen the RDFa community name  
drop first Yahoo! and then Google as evidence of success of the design  
of RDFa. However, in both cases when others inspected whether the  
implementations actually implemented CURIE processing, the answer was  
that they didn't. Yahoo! first reportedly treated prefixes as  
meaningful (http://lists.w3.org/Archives/Public/public-rdf-in-xhtml-tf/2009Mar/0100.html 
). Google reportedly looks at the after-colon part, and a Google  
engineer indicated that the implementation "will [...] deviate from  
the standard" (http://lists.w3.org/Archives/Public/public-rdf-in-xhtml-tf/2009Sep/0126.html 

I see both Yahoo!'s case and Google's case as evidence suggesting  
CURIEs are a design that isn't working.

Now, when Drupal is name dropped as evidence of success, why should  
anyone trust that it actually implements RDFa either in producer or  
consumer capacity in a way that in any meaningful sense validates the  
RDFa design until some further details about the matter are presented?

> Do you know how big the Drupal community is?

I don't. How big is it?

> it's becoming farcical.

How would you characterize the ongoing denial that the syntax xmlns:p="http://example.com/ 
" is problematic?

How can the problem be meaningfully resolved when you aren't even  
admitting there's a problem to discuss?

> If you or Ian had proposed that RDFa also support reverse DNS
> identifiers, you might have found less support...but hey, let's talk
> about it.
> But that didn't happen.

I did suggest using full URIs instead of CURIEs. I even prototyped  
validator support for it before we had Microdata.

I removed the validator prototype when it became obvious that the RDFa  
community was utterly disinterested in solving the xmlns problem by  
using full URIs and when a better alternative (Microdata) had been  

Henri Sivonen

Received on Tuesday, 22 September 2009 20:17:08 UTC