Re: JAWS and SVG ARIA markup

On Wed, 10 Feb 2016 18:00:51 +0100, Fred Esch <fesch@us.ibm.com> wrote:

> I used a couple of files that are generated from the testable statements  
> in the wiki (zip of files sent on the thread with Chaals).

That message is  
http://www.w3.org/mid/201602082104.u18L44xf023344@d01av02.pok.ibm.com

But the attachments weren't recognised by the W3C archive system -  
something weird in Fred's email? - so unless you want to extract the  
encoded text, ask Fred or me for a copy.

> Using Firefox with JAWS I could hear the markup being read. It appears
> that JAWS identifies, aria-labeledby, aria-label, title and contents
> of a text element. The tests are simple so I don't know what it does if
> there are multiple elements with markup. I expect it picks the first one.

For more complex things, it seems that most browsers just pick the first  
title element in the document and present that as the document's title.

While most browsers will present title elements as a tooltip on hovering  
with a mouse, it seems VoiceOver never gets anything except the first one  
 from the browsers I tested on Mac.

> In trying 1 file with NVDA, it read the markup too.

It would be good if you had a look at e.g.  
http://svg-access-w3cg.github.io/use-case-examples/rectrack2-notes.html  
which provides more information about a number of browsers, interacting  
with a more complex "real-world" SVG - and ideally added your observations  
for Windows browsers. At the moment it seems that some of what you are  
doing is great, and some of it is just repeating what has already been  
done and recorded at  
http://svg-access-w3cg.github.io/svg-a11y-tests/index.html

We should figure out how to track what we already have better, instead of  
repeatedly doing important but boring work and losing the results.

> Not a conclusive test, but it suggests that we should recommend marking
> up inline SVG rather than use figure and figcaption.

I don't think recommending inline SVG and figure/figcaption are in  
conflict, and I think it would be good if people did both. Unfortunately  
there is no neat way to point to a structured description from an inline  
SVG, so where that is required the link should be part of the caption.

It is important for users whose browser doesn't handle SVG accessibility  
very well - which means more or less everyone in practice, unless the SVG  
is *very* carefully constructed - whenever there is more to the  
description than is a comfortable caption. Which is probably a high  
proportion of cases where SVG is used for e.g. auto-generated charts and  
the like.

cheers

Chaals

-- 
Charles McCathie Nevile - web standards - CTO Office, Yandex
  chaals@yandex-team.ru - - - Find more at http://yandex.com

Received on Thursday, 11 February 2016 11:02:30 UTC