Re: Next steps for the ARIA syntax discussion

Hi Charles,

On Jun 3, 2008, at 7:27 PM, Charles McCathieNevile wrote:

> On Tue, 03 Jun 2008 08:17:17 -0300, Robert J Burns  
> <rob@robburns.com> wrote:
>
>> Hi Charles,
>>
>> On Jun 1, 2008, at 3:01 AM, Charles McCathieNevile wrote:
>
>>> The mistake was suggesting that there should be a namespace *other  
>>> than the null string* for aria attributes...
>
>> The XML namespaces recommendation is designed to be modular. The  
>> only thing other recommendations need to concern themselves with is  
>> providing a namespace URI to uniquely identify the vocabulary.
>
> Indeed. And in order to maintain compatibility with the Web as it  
> has developed for the last decade and a half, in the case of  
> attributes the default case of having a null namespace ("in no  
> namespace", "blank namespace", call it what you will) is the only  
> solution that works for HTML and XHTML serialisations. Fortunately  
> that is the solution compatible with all the relevant specs, and  
> with all the deployed technology.

This is a stretch. Yes it is compatible. However, its irrelevant to  
the discussion of HTML5. Whether it is fortunate or not is a  
completely off-topic issue (unless we start talking about the  
introduction of namespace to text/html).

>
>
>> As for the text/html serialization we should be careful here as  
>> well. As TAG has already expressed a desire to add distributed  
>> extensibility to text/html in the future, we should again be  
>> careful here. I don't see a problem with the no namespace name (or  
>> null nameespace since the difference in what we call it here is  
>> trivial).  However, it is hard to foresee what problems we might  
>> cause for W3C if we start moving HTML5 specified elements and  
>> attributes into other various namespaces (other than the html  
>> namespace and the null namespace or no name namespace).
>
> It appears we agree. The elements should be in the html namespace.  
> The attributes in the null namespace.

But only for non-XML namespace aware documents (or by extension  
documents unaware of other namespace recommendations). The point I'm  
trying to make and that I'm not sure we agree on is that  
recommendations such as HTML5 should not be trying to anticipate what  
the XML namespaces recommendation says about their attributes. It  
creates dependencies and undermines valuable abstractions, and we  
don't want to go there.

If that all makes sense to you and you agree, then indeed we do agree.

Take care,
Rob

Received on Tuesday, 3 June 2008 19:32:45 UTC