Re: HREF attribute in elements other than A and LINK

For the record, I don't really mind one way or the other.  It is on 
every element in the xhtml-rdfa draft [1], DTD and XML Schema now 
because I was asked to migrate the modules from working in XHTML M12N 2 
to working in XHTML M12N 1.1.  There is nothing in the current work 
that, as far as I know, requires that @href and its associated other 
attributes from XHTML 2 be incorporated into RDFa in an XHTML 1.1 
context.  If you all tell me that it is not required and to change it, 
great.  If not, also great.

While we are looking at this, though, I think we need to look at meta 
and link.  Again, in migrating from XHTML M12N 2, I have changed the 
content model of they are available in the body and so that link has a 
content model of (link | meta)* and meta has a content model of (PCDATA 
| Inline)*.  This is clearly a change from XHTML 1.1.

[1]  http://www.w3.org/MarkUp/Drafts/xhtml-rdfa

Ivan Herman wrote:
> +1 with all your arguments!
>
> Ivan
>
> Ben Adida wrote:
>   
>> (Chair hat off)
>>
>> Mark Birbeck wrote:
>>
>> [...]
>>
>>     
>>> My point about starting with a 'mental model' is only to suggest that
>>> we look at how something fits with existing HTML, and existing
>>> practice. The idea was simply that if we're happy with the model we
>>> can look at how we explain it to other people. (It's actually how
>>> nearly everything in RDFa has been designed, just not always
>>> explicitly.)
>>>       
>> So, I'm not happy with the model of adding HREF everywhere in something
>> called XHTML1.1+RDFa. Mark, you've always pointed out that RDFa is
>> primarily aimed at HTML authors adding semantic markup. I mostly agree
>> with you. With that assumption, I can only imagine an HTML author being
>> thoroughly confused by HREF everywhere, wondering, "where's the
>> clickable link? How do I *make* it clickable?". Why would an HTML author
>> do this? What is the use case that would lead him to put an HREF on a
>> DIV as far as the HTML author is concerned? This adds a whole level of
>> inherently invisible metadata, where the primary goal of RDFa is to mark
>> up visible data.
>>
>> Regarding acceptability of this approach, I'm in full agreement with
>> Ivan on this: the backlash against this will be enormous. We *have* to
>> plan for it, and, more importantly, we have to ask ourselves: what is
>> the cost/benefit of this quasi-ensured backlash? I see a high cost, and
>> I don't see the benefit wrt our goals.
>>
>> Finally, the biggest worry I have is regarding the perception of this
>> change. If we add HREF in a bunch of places, we're really changing the
>> document model for HTML in ways that even adding REL didn't do (since
>> that is still about marking up visible content). It's not XHTML1.1
>> anymore. It's clearly XHTML1.2. And the perception will be that we're
>> trying to squeeze XHTML2 features into XHTML1 via RDFa. That is a
>> dangerous proposition: we should not make RDFa an even bigger lightning
>> bolt for criticism, if we can help it.
>>
>> -Ben
>>     
>
>   

-- 
Shane P. McCarron                          Phone: +1 763 786-8160 x120
Managing Director                            Fax: +1 763 786-8180
ApTest Minnesota                            Inet: shane@aptest.com

Received on Tuesday, 17 April 2007 13:01:16 UTC