Re: ISSUE-82 - profile-disambiguation - Chairs Solicit Proposals

On Feb 25, 2010, at 6:42 AM, Julian Reschke wrote:

> On 24.02.2010 10:18, Maciej Stachowiak wrote:
>> ...
>> I would personally recommend option #2. I think if we are going to  
>> have
>> a separate profile spec, then it there's no point mucking with the
>> profile definition in the HTML5. The separate profile draft would
>> automatically make this use of profile conforming in text/html so  
>> long
>> as the profile draft is considered an applicable specification,  
>> whatever
>> our IANA registration says. And changing the details of HTML5's
>> processing requirements would not change. If anything, we should
>> probably remove most of what HTML5 says about @profile currently.
>> ...
>
>
> This is simply incorrect. Assuming we *do* publish a separate spec  
> defining @profile, it would remain incorrect and still would need to  
> be fixed.
>
> Is the "applicable specification" rule for extensions meant to allow  
> extensions to override "processing requirements"?

My interpretation is yes. Current examples in proposed extension drafts:

* HTML+RDFa overrides the processing requirements and extends the  
syntax for the @rel and @rev attributes.
* HTML Microdata extends the processing requirements for the <time>  
element and its datetime attribute, the <link> element, and the <meta>  
element. For <link> and <meta>, it also changes the contexts in which  
they are allowed to appear.
* HTML Canvas 2D Context modifies the processing requirements for the  
getContext() method on HTMLCanvasElement.
* WebGL (a Khronos spec, not being developed in this WG) also modifies  
the processing requirements for the getContext() method on  
HTMLCanvasElement.

So it seems generally accepted that extension specs may override or  
modify processing requirements for existing elements, attributes and  
DOM APIs in the HTML5 draft.

On the telecon, you volunteered to write a Change Proposal for  
ISSUE-82, with a deadline of a month from today. I will record that on  
the issue status list. If you conclude that it is not necessary after  
all to change HTML5's definition of the profile attribute for this  
issue, and that the profile extension spec is in fact sufficient, then  
please let us know.

Regards,
Maciej

Received on Thursday, 25 February 2010 17:52:35 UTC