Re: False aria-describedby expectations in ARIA Authoring Practices (longdesc)

HI Maciej,

> The interaction model doesn't have to be that of a link to still reflect the semantics of the referenced text. A screen reader could simply read the referenced >content when the user chooses to see a description, including all the usual cues it would if reading that markup directly, but then return you to where you were in >the document. I tend to think that would be a better user experience than a model where you follow a link and then navigate back.


I agree with this, but how is this to be implemented and where? in the
browser or in the screen reader?

we have at least one vendor (NVDA)  who has indicated they would not
implement  behaviour in their product to provide access to longdesc
(which is essentially a element specific version of what we are
talking about for describedby), but would support it if the UI was
exposed via browsers.

One of the issues with longdesc has been that it is not exposed by
browsers (other than the url string in the DOM). I consider that the
best chnace we have of making aria-describedby and/or longdesc useful
and usable is for the UI to be in the browser.

regards
stevef

On 23 April 2011 09:31, Maciej Stachowiak <mjs@apple.com> wrote:
>
> On Apr 22, 2011, at 11:13 PM, Jonas Sicking wrote:
>
>> On Fri, Apr 22, 2011 at 6:12 PM, Leif Halvard Silli
>> <xn--mlform-iua@xn--mlform-iua.no> wrote:
>>> Jonas Sicking, Fri, 22 Apr 2011 15:08:11 -0700:
>>>> On Fri, Apr 22, 2011 at 2:25 PM, Steve Faulkner wrote:
>>>
>>>>> In fact, we could even display a context menu item
>>>> for each link if there are multiple as aria-describedby allows
>>>> pointing to multiple elements, many of which could be links.
>>>
>>> Then some would use @aria-describedby for the purpose of the link,
>>> while others would use it for the purpose of its text content = purpose
>>> crash.
>>
>> That is your interpretation of the ARIA spec. One that I don't agree
>> with. And based on the subject of this thread, one that the spec
>> editors might not agree with either.
>>
>> My interpretation is that everyone should see the link. *That* would
>> be the best solution for everyone. No matter what we do with regards
>> to @longdesc it seems that it's better for AT users if they do indeed
>> see the semantics of the elements pointed to by aria-describedby.
>
>
> I'm not sure if you have in mind the semantics of a link, or the user experience of a link.
>
> The interaction model doesn't have to be that of a link to still reflect the semantics of the referenced text. A screen reader could simply read the referenced content when the user chooses to see a description, including all the usual cues it would if reading that markup directly, but then return you to where you were in the document. I tend to think that would be a better user experience than a model where you follow a link and then navigate back.
>
> Regards,
> Maciej
>
>



-- 
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, 23 April 2011 09:11:47 UTC