Re: img@relaxed CP [was: CfC: Close ISSUE-206: meta-generator by Amicable Resolution]

Hi ben,

if you have been following the discussions, in the various fora, it
should be clear that this attribute is not meant to be added ad hoc by
downstream developers, it is designed to be automatically added by
content generators in very limited circumstances i.e. when the
sofwtare emits an image that is known to have meaning, but that
meaning is unknown at the time of publication.

If it is designed and promoted as a means to stop validators whining
about ANY image without alt, whether it be a spacer gif or an image
containing complex information, then I do not believe it will gain the
required consensus within the working group.

> We've already got negative signals to indicate that (don't include
> alt="", don't include role="presentation", don't include
> aria-hidden="true", etc.) We don't need a positive signal too,
> especially as it wouldn't solve the problem for ingested content as
> there would be no sensible way to apply it.


We also have another negative siganl which is at the root of the issue
the lack of alt. In cases where there is lack of alt attribute, in
conforming HTML5 documents  this is supposed to signify that the image
is a key part of the content which should have a text alternative but
one is not available at the time of publication. The allowing of a
validator flag is built upon this concept.

So either the concept is flawed and should be removed from HTML5 or
the concept is sound and the addition of the attribute will provide
such indication unabiguously.

If the concept is flawed, then we must reject the notion of permitting
a signifier and go back to not providing a method for automated
content generators from opting out of validation via a flag on images
of whatever flavour.


regards
SteveF



On 4 August 2012 10:40, Benjamin Hawkes-Lewis
<bhawkeslewis@googlemail.com> wrote:
> On Sat, Aug 4, 2012 at 5:04 AM, Steve Faulkner <faulkner.steve@gmail.com> wrote:
>>> When a developer knows an image does not contain important content,
>>> they can use alt="" and it will not be exposed to the accessibility
>>> hierarchy.
>>>
>>> Equally when a developer does not know the alternative text for an
>>> image, they can omit alt and it *will* be exposed to the accessibility
>>> hierarchy.
>>
>> you are making a false assumption, this only works if the generator
>> image developer controls where the content is published.
>>  The developer does not necessarily control the content of the page
>> the image is embedded in, so the developer cannot ensure that only
>> images that are significant have alt omitted.
>
> The developer cannot add an additional attribute only to images that
> *are* significant in ingested third-party content any more than they
> can provide appropriate text equivalents.
>
> I don't think it would be a good idea for developers to ingest images
> in third-party content with alt="", assume they nevertheless _might_
> be significant, and override those provided text alternatives that
> hide the images by robotically adding an attribute to force exposure
> of all images in third-party content to the accessibility hierarchy,
> if that's what you're envisaging (?).
>
>> that is not the scope of the discussion,  it has been identified that
>> providing an indication of the image significance to AT is an
>> important aspect for some involved
>
> We've already got negative signals to indicate that (don't include
> alt="", don't include role="presentation", don't include
> aria-hidden="true", etc.) We don't need a positive signal too,
> especially as it wouldn't solve the problem for ingested content as
> there would be no sensible way to apply it.
>
> --
> Benjamin Hawkes-Lewis



-- 
with regards

Steve Faulkner
Technical Director - TPG

www.paciellogroup.com | www.HTML5accessibility.com |
www.twitter.com/stevefaulkner
HTML5: Techniques for providing useful text alternatives -
dev.w3.org/html5/alt-techniques/
Web Accessibility Toolbar - www.paciellogroup.com/resources/wat-ie-about.html

Received on Saturday, 4 August 2012 10:23:06 UTC