Re: minutes for HTML WG f2f, 2011-11-04, part 1

On Sun, 06 Nov 2011 09:46:08 +0100, Peter Winnberg  
<peter.winnberg@gmail.com> wrote:

> 2011/11/6 Sam Ruby <rubys@intertwingly.net>:
>> I will note that there is one concern that has been identified with the
>> definition of the data element as currently specified:
>>
>> http://www.w3.org/Bugs/Public/show_bug.cgi?id=13240#c72
>>
>> See the last paragraph in that comment.  Given the sentiment in the  
>> room in
>> the F2F, if that concern can be addressed and no other concerns are  
>> voiced,
>> it occurs to me that we could move this forward quickly with a Call for
>> Consensus.  Should anybody object, we would give those who do an  
>> opportunity
>> to produce counter proposals.
>
> The problem that I have with the data element is that I have not seen
> anyone explain is why the data element would be a superior solution to
> say:
>
> <span value="machine-readable data here">Data here</span>
>
> Why is the data element needed when it seems to have the same
> semantics as span? If I see some convincing examples of why this is
> needed I might redraw my objections to it. Otherwise yes, it is likely
> that I would do a counter proposal.
>
> The way I see it, only a attribute is needed to hold the
> machine-readable data, and this attribute could perhaps be used on
> other elements as well (time?). If something like that would be done
> instead, and not add a new element and just an attribute, there are of
> course other things to consider, what should that attribute be named,
> and how can it be specified so that microdata/microformats/RDFa could
> take advantage of it.

The main problem with a global attribute is that it would make it less  
clear which attribute takes precedence and whether or not the value is  
resolved as a URL, as previously discussed. [1][2] Further, what would the  
global attribute be? value="" already exists with different semantics on  
<button>, <option>, <input>, <li>, <meter>, <progress> and <param>.  
content="" is not an option since RDFa uses it (in the early days  
microdata had both a property="" and an about="" attribute and there were  
objections to this).

[1] http://www.w3.org/Bugs/Public/show_bug.cgi?id=13240#c17
[2] http://lists.w3.org/Archives/Public/public-html/2011Oct/0166.html

-- 
Philip Jägenstedt
Core Developer
Opera Software

Received on Tuesday, 8 November 2011 10:05:48 UTC