W3C home > Mailing lists > Public > w3c-wai-gl@w3.org > October to December 2013

Re: Re: Clarification of WCAG intent and meaning of techniques [Re: WCAG considering amending F65 to NOT fail missing ALT text if title or aria-label is present]

From: Leif Halvard Silli <xn--mlform-iua@xn--mlform-iua.no>
Date: Thu, 28 Nov 2013 21:12:18 +0100
To: Loretta Guarino Reid <lorettaguarino@google.com>
Cc: Steve Faulkner <faulkner.steve@gmail.com>, "Schnabel, Stefan" <stefan.schnabel@sap.com>, Michael Cooper <cooper@w3.org>, David MacDonald <david100@sympatico.ca>, Janina Sajka <janina@rednote.net>, HTML Accessibility Task Force <public-html-a11y@w3.org>, WCAG WG <w3c-wai-gl@w3.org>, "public-comments-wcag20@w3.org" <public-comments-wcag20@w3.org>, Gregg Vanderheiden <gv@trace.wisc.edu>, "kirsten@can-adapt.com" <kirsten@can-adapt.com>
Message-ID: <20131128211218829557.a9255635@xn--mlform-iua.no>
Loretta Guarino Reid, Thu, 28 Nov 2013 11:02:54 -0800:
> I disagree. This is an example of using text already on the page as the
> image description.

Agreed.

> This is not an attempt to use aria-labelled by as an
> alternate implementation of alt.

From the most literal perspective, you are right. Because, after all, 
the example uses aria-labelLEDBY - and not aria-label. However, 
semantically, aria-label and aria-labelledby are equivalents. They only 
differ in where they take the label text from: One is a label 
container, the other points to the label container(s). Therefore, from 
a semantic perspective, you are wrong: It doesn't matter whether you 
use aria-label or aria-labelledby. From this semantic view, it makes 
perfect sense that aria-labelledby does not prevent redundancy.

> It is a use case where alt is redundant if
> the image description is already present and can be associated with the
> image via aria-labelledby.

Avoid redundancy is precisely the role of aria-hidden="true" - that's 
why I added it on the <p> element.

What we are looking for is a *programmatic* association of image and 
text.[0][1][2] In HTML5, the use of <figure> plus <figcaption> provides 
such an association, without use of aria-labelledby. That's why HTML5 
allows @alt to be omitted on <img> when it occurs inside <figure> with 
a <figcaption>. If you were to replace <figure> with a <div>, you would 
need to use aria-labelledby in order to programmatically associate a 
child <span id=label> element with the img. But then you would also get 
that redundancy problem.

PS: The aria-labelledby can point to more than one element. And in such 
a situation, it might make perfect sense to *not* use 
aria-hidden="true" to hide the adjacent element *even if* it this means 
that *one* part label (namely the text of the adjacent element) is 
repeated.

[0] http://www.w3.org/TR/WCAG-TECHS/F65.html#F65-applicability

[1] http://www.w3.org/TR/2008/REC-WCAG20-20081211/#text-equiv-all

[2] http://www.w3.org/TR/2008/REC-WCAG20-20081211/#text-altdef


Leif H Silli

> On Thu, Nov 28, 2013 at 9:51 AM, Leif H. Silli wrote:
> 
>>  Loretta Guarino Reid, 28.11.2013, 17:45
>> 
>>> Adding aria-hidden completely changes this example, which is one where
>> the
>>> image description is already available as *visible* text. This is a
>> similar
>>> use case to figure and figcaption.
>> 
>> Well, it is the aria-labelledby that “completely changes" the example.
>> Aria-hidden merely deals with the effect of the aria-labelledby.
>> 
>> Leif Halvard Silli
>> 
>> 
Received on Thursday, 28 November 2013 20:12:54 UTC

This archive was generated by hypermail 2.4.0 : Friday, 17 January 2020 20:32:54 UTC