W3C home > Mailing lists > Public > public-html@w3.org > May 2009

Re: use case focus - resending

From: Sam Ruby <rubys@intertwingly.net>
Date: Tue, 12 May 2009 19:08:28 -0400
Message-ID: <4A0A016C.1000904@intertwingly.net>
To: Shelley Powers <shelleyp@burningbird.net>
CC: public-html@w3.org
Shelley Powers wrote:
> Sam Ruby wrote:
>> Shelley Powers wrote:
>>>
>>>>> After today, though, I'm less concerned. Much less concerned.
>>>>
>>>> :-)
>>>>
>>>> As near as I can tell, the announcement is totally vaporware at this 
>>>> point, but at least it is from a credible source.
>>>>
>>> The RDFa support is also constrained now, to more or less match 
>>> microformat support.  Google created its own vocabulary, which is 
>>> cool. People can use more than one.
>>>
>>> The company has said that it will _gradually_ add in support for 
>>> other vocabularies over time. It's not a use what you want 
>>> annotation, at the moment, but I have hopes for over time.
>>>
>>>> I still maintain that somebody needs to produce an "RDFa for HTML" 
>>>> draft specification; the question as to whether that particular 
>>>> content needs to be in or separate from the HTML 5 specification are 
>>>> secondary, IMHO.
>>>>
>>> I agree, and I believe this is happening. But I think this happening 
>>> is where some of the contention arose.
>>
>> If you can find an URI for such a document, I'd appreciate it.  I've 
>> been asking and I've been looking, and I can't find it.
> 
> Not a document, this was from emails and WhatWG listing posts where RDFa 
> folks and WhatWG folks seemed to be heading to an accord on how RDFa 
> could possibly be incorporated into HTML. I also have the impression 
> this was happening with the RDFa folks, but I'm not a member of that 
> working group.
>>
>>> If a person doesn't care about validation, one can use RDFa with few 
>>> worries.
>>
>> If you use the DOM and know the difference between a node name and a 
>> local name, the differences between HTML parsing rules and XHTML 
>> parsing rules are fairly minor.  If you use SAX or XOM, the 
>> differences are a bit larger.  Still finite, but definitely larger.  
>> (Ignore Henri when he points out that the number of xmlns:foo 
>> attributes are "countably infinite", so too are data attributes).
>>
>> If the right solution for this is to replace the usage of xmlns with 
>> prefix, the cost for this change goes up with every application 
>> (producer or consumer) that implements RDFa as it is spec'ed now.
>>
> You have to judge against what you get in return: being able to 
> incorporate RDFa validly into HTML is a powerful incentive. And perhaps 
> this approach won't be used, I don't know.
> 
>> I also believe that Philip can produce a number of examples which 
>> would indicate that "few worries" is a bit of an understatement.  Then 
>> again, if anybody could build a solid case that the humble <p> element 
>> in HTML4 is unworkable, I'm sure that person would be Philip.
> 
> For the average person, there are not that many worries...if you don't 
> care about validation. Or you use XHTML 1.1, but served up as HTML, 
> which is what Drupal is doing.

... which every browser on the planet treats as HTML.  I say this as a 
person who fairly consistently serves up well-formed XHTML properly to a 
person who also fairly consistently serves up well-formed XHTML properly.

> The thing is, RDFa is happening now, and just got a boost from Google, 
> and will get another one with Drupal 7. It would be nice if it play 
> together with the HTML of the future.

The consequences of "happening now" based on where people "seem to be 
heading" without "caring about validation" is that people will get 
things like prefixes wrong:

http://lists.w3.org/Archives/Public/public-rdf-in-xhtml-tf/2009May/0064.html

And things like namespaces wrong:

http://lists.whatwg.org/htdig.cgi/whatwg-whatwg.org/2009-May/019717.html

The inevitable next step is that consuming libraries start 
"compensating" for common errors.  I don't think anybody here wants to 
go there.

If a spec is produced, and people who care about conformance and interop 
produce non-artificial content, and there are real applications for this 
data, then I will do what I can to ensure that there is a validator for 
this.

As far as I can tell, all of the preconditions I gave in the preceding 
paragraph are met... save one.

> Shelley

- Sam Ruby
Received on Tuesday, 12 May 2009 23:09:12 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Wednesday, 9 May 2012 00:16:36 GMT