Re: Questions about the new "caption" role for tables

On 03/31/2015 12:43 PM, Joanmarie Diggs wrote:
> Hey Steve.
> 
> On 03/31/2015 11:36 AM, Steve Faulkner wrote:
>> "In addition, at least in the case of HTML, there is an associated
>> positional requirement, namely that the caption (if present) should be
>> the first child element" 
>>
>> It is my understanding  that the positional requirement is due to parsing implementation constraints only. A suggested constraint to have is that each role=table or role=group (in the case of figcaption) only have one child with a role=caption. 
> 
> I agree with the constraint and will add text to that effect. Thanks!
> 
> As for the positional requirement and parsing, I don't think the
> aforementioned constraint solves that. User agents still have to parse
> tables. Only now, with the creation of the new table roles, we're
> empowering authors to create tables out of divs which strikes me as the
> same parsing problem, with an increased likelihood of authoring and
> parsing errors, and increased difficulty in validating the result
> programmatically.
> 
> Hmmmm.... I've just convinced myself that we should indeed have a
> positional requirement. ;) I'm willing to hear arguments to the
> contrary, however.

Hmmmm again: That positional requirement works if the element with
role="caption" is actually contained within the element it applies to.
But if instead it's owned by the element it applies to, we can't do
that. (Right?)

Thoughts/feedback/insight <strike>welcome</strike> strongly encouraged.
--joanie

Received on Tuesday, 31 March 2015 18:51:00 UTC