Re: [BULK] - Re: [XHTML2] Spirit of "1.1.3. XHTML 2 and Presentation" (PR#7759)

At first, Ian's request seems perfectly reasonable, and succinctly put.

Upon closer examination, there is not such a tension between the text of 
1.1.3 and the presence of the style attribute, any more than there is a 
tension between the text of 1.1.3 and the <style> element. Although 
section 1.1.3. does indicate that XHTML 2 is "removing all presentation 
elements", the style attribute is not a presentation element. It's not 
even a presentation attribute, in the sense that "fontSize" or 
"lineSpacing" would be presentation attributes. In other words, the 
style attribute's *value* contains stylesheet information bound to a 
single element, much in the same way that the <style> element's textual 
content contains stylesheet information mapped to individual elements 
within the document. Presentation attributes, in my opinion, would be 
those in which the attribute name represented a style predicate, with 
the attribute value representing the value of that style predicate.

The style attribute, on the other hand, contains a value that consists 
of many presentation predicates and their values. The style attribute is 
no more than locally-bound information transferred from the <style> 
element; hence it is consistent with the stated goal of 1.1.3 of 
"subordinating all presentation to style sheets", as the style attribute 
content is really a mini-stylesheet.

Ian does not seem to be making the same request regarding the <style> 
element, and as such his request would seem more inconsistent than the 
tension between 1.1.3 and the presence of the style attribute.

On the other hand, this question is likely to come up more in the 
future, and the working group ignores it now at their peril---putting 
text in the recommendation would be preemptive against answer the 
question again and again after the document is finalized.

Lastly, I'd like to personally express my desire to retain the style 
attribute. With newer AJAX applications, it is many times extremely 
useful, if not necessary, to send style information to a browser for 
small updated within a document. The lack of a style attribute would 
mean that dynamically updating the style of a single element within a 
document would require somehow modifying an external stylesheet or the 
internal stylesheet in the document's <head>.

Garret

Ian Hickson wrote:
> On Wed, 8 Feb 2006, Epperson, Beth wrote:
>   
>> I would suggest that this be a response from the CSS WG. As a member of 
>> the WG - Ian, I would suggest you bring this to the F2F where the CSS WG 
>> come to an internal agreement as to whether the desire is to have the 
>> attribute removed or retained. Once the CSS WG has a decision, then 
>> request a joint meeting where both WG's can discuss and finally agree 
>> and be done with this discussion. But wait - we already did that 2 or 3 
>> years ago - hence it is in the spec.
>>     
>
> In that case, as I said in my original request:
>
>   
>>>>>> XHTML2's "1.1.3. XHTML 2 and Presentation" section says:
>>>>>>
>>>>>> # XHTML 2 takes HTML back to these roots, by removing all presentation
>>>>>> # elements, and subordinating all presentation to style sheets.
>>>>>>
>>>>>> This, while technically true, does not seem completely 
>>>>>> consistent with the inclusion of the style="" attribute (in 
>>>>>> section 27 XHTML Style Attribute Module). I feel that the intent 
>>>>>> of removing all presentational aspects from the language is to 
>>>>>> be applauded and would like to ask for the purely presentational 
>>>>>> style="" attribute to be removed.
>>>>>>
>>>>>> If it is to remain in XHTML2, however, a clear indication of its 
>>>>>> purpose should be given, so as to explain the conflict between 
>>>>>> its presence and the sentiment of section 1.1.3.
>>>>>>             
>
> What I am asking for is for EITHER:
>
>  A. The "style" attribute to be removed, to be consistent with the 
>     sentiment in section 1.1.3,
>
> OR:
>
>  B. The text in section 1.1.3 to explain why it is considered ok to have 
>     presentational attributes but not presentational elements.
>
> That's all I'm asking for. It should not be difficult to do option B, if 
> there really is a good reason not to do option A. Just say what the reason 
> is in section 1.1.3.
>
>

Received on Thursday, 9 February 2006 23:02:10 UTC