Re: ISSUE-30 (Longdesc): WebAIM 2009 survey

Steve Axthelm On 09-11-02 17.42:

> WebAIM released the results of their 2009 screenreader survey. 
> One of the questions I thought relevant to this issue:
> 
> <http://webaim.org/projects/screenreadersurvey2/#images>
> 
>      "Some images, such as charts, diagrams, or comic strips, 
> are too
>      complex to describe in only a few words. If a long, detailed
>      description of these images is available, how would you 
> prefer to
>      have it presented to you?"
> 
> 
>      "There is no clear consensus in these responses. However, the
>      in-page options outweigh the options that place the longer
>      description on another page. Interestingly, the option of placing
>      the alternative on a separate page but having it announced 
> by the
>      screen reader, the current behavior of images with the longdesc
>      attribute, was a very unpopular option, second only to 
> being ignored
>      entirely."

Thanks for notification about an interesting and relevant survey. 
Personally, I see the results as quite supportive of the way I see 
@longdesc.

One thing to clarify first: a @longdesc does not per definition 
need point to another page. It could very well point to an element 
that follows directly after the image. I think we need to get this 
into both HTML 5 and ARIA!

I miss a clear answer in your survey, on whether AT users also 
would have liked a way to "jump over" the the long description 
when/if they are satisfied with the short one. For example, one 
thing is that they can jump from the short/the image to a relevant 
  description. However, those who prefer the long description to 
immediately follow the image/the short description, would they not 
like to know/be notified that this is the "this is just the long 
fallback for the image - hence you may skip if need be" ?

An evaluation of the results:

"On a separate page, available by following a link" together with 
"On a separate page, announced by and available to my screen 
reader" are very similar options, in my view, even if it is 
interesting that your readers preferred the least automatic 
variant - the link (and @longdesc is a link). Together they count 
for 181 voices, though, and thus become the largest group. These 
users effectively like to not be disturbed in their reading by the 
long description itself. I imagine that if some element is marked 
as "long description", and if the UA doesn't read such text unless 
the user asks it to do so, then it would not matter to these 
readers whether the description is in the page itself or on a 
separate page?

"As text on the web page, immediately following the image" is the 
largest single group. It seems to be an answer that can be 
interpreted in many ways. Their motivation for this answer could 
be that they would like to get the long description without having 
to wait. It could also be motivated by normal expectations of a 
text: things that belong together should be close to one another. 
This has to do both with trusting the text and many other things. 
Also, a possible motivation for saying this could perhaps be that 
they prefer that the page looks the same for both blind and 
seeing. What is not clear to me, is whether those who answered 
like this would be against a way to jump over/ignore = identify 
the long description, both where it starts and where it ends ... I 
  suppose not.

The second highest score, "As optional text, available on the same 
page but only if I request it by following a link", could be said 
to cover the way I say that @longdesc can be used: to point to a 
place on the same page. However, I suppose that those who answered 
like this did not assume that the text *has* to be far away from 
the image. I suppose that they simply expect to activate the 
description by a simple click. The text could be close to the 
image, though. Thus their wish would not be very far from the 
"winning answer" - "As text on the web page, immediately following 
the image".

My own thoughts: It seems to me that it would not have been so bad 
to have a specific "fallback element" - like <alt></alt>. Such a 
thing would be easy to ignore/identify, if I am right in that such 
ignoring/identification would be a good thjing. Such elements 
could could follow immediately after the IMG *and* it could also 
be accessible via a link (longdesc). (Sorry about the lack of 
pointer, but <alt></alt> has been suggested in this group earlier 
by others than me.)

The advantage of <alt> over both aria-describedby and longdesc, is 
that it doesn't rely a link or idref pointer in another element - 
the element itself makes itself known as 
fallback/description/equivalent.

The name of the element did not need to be <alt>. It could be 
something more telling, like <textequiv> or just <equiv> (for 
<equivalent>). Such elements did not need to always be hidden for 
sighted users, since sighted users may also need to identify a 
text as equivalent of a specific image.

And in this regard: another thing I miss in your survey is that 
you had taken in the existence of <object> ... When <object> 
contains a static image, then inside that <object> is were the 
long description follows. Thus, when <object> is used, then the 
reader have every option to just jump over it, as the fallback 
have a clear beginning and a clear end.

The only drawback I see w.r.t. use of <object> is that it 
currently doesn't offer any method for separating short and long 
fallback.
-- 
leif halvard silli

Received on Monday, 2 November 2009 18:37:14 UTC