Comments on: Screen Reader Support for ARIA Live Regions

By: Tink

Tink — Sun, 28 Jun 2009 21:46:22 +0000

Updated the post to include results from Hal v11.

By: Tink

Tink — Tue, 09 Jun 2009 18:19:42 +0000

For the moment, my Hal demo has run out. I often wish that Dolphin opted for a 30/40 minute demo in the style of Jaws or Window Eyes, it would make testing during development much easier.

I’d discounted Thunder because of its reliance on WebbIE, which didn’t support JavaScript/AJAX last time I checked… Looking at the blurb, this still seems to be true, but I’ll download a copy and give it a whirl all the same.

If you’re able to contribute any results from Hal, I’d be glad to include them here though. Thanks.

By: Steve Griffiths

Steve Griffiths — Mon, 08 Jun 2009 11:04:54 +0000

Only just come across this. Was there any particular reason why Hal and Thunder weren’t included in the test?

By: Tink

Tink — Tue, 02 Jun 2009 18:11:38 +0000

Radek, what makes this really interesting is that now we’ve got a situation where some screen readers support a property value that’s no longer part of the spec.

By: Tink

Tink — Tue, 02 Jun 2009 18:03:16 +0000

Thanks James & Lloyd. I’ve updated the screen reader versions in the tables above.

By: Radek Pavlicek

Radek Pavlicek — Tue, 02 Jun 2009 14:40:15 +0000

I think that live=”rude” isn´t already part of WAI-ARIA specification, see http://www.w3.org/WAI/PF/aria/#aria-live. Live has only three values – off, polite, assertive.

By: Lloyd Rasmussen

Lloyd Rasmussen — Tue, 02 Jun 2009 14:26:25 +0000

In the case of Window-Eyes, is this a beta of 7.1 or is it 7.01, the current version? I realize that the user wouldn’t know that an update had occured, but I assume that if she does Insert-backslash to refresh the screen (which refreshes the browse buffer) the information would get updated. I like the names for these types of regions, and hope that “rude” regions don’t get created very often.

By: James Teh

James Teh — Mon, 01 Jun 2009 22:41:20 +0000

NVDA does support additions for polite and rude live regions for “show” events. For example, if a container is marked as live and a div is appended to the container or a div inside the container becomes visible, it will be announced. Assertive live regions are currently treated like polite live regions. Obviously, this is a grievously incomplete implementation and we do have plans to fix it, but I thought it worth noting that we do have some support nevertheless.

Btw, to avoid confusion, it’d be great if you could be more specific about the version of NVDA used. NVDA 0.6 does not yet exist. The latest official release is NVDA 0.6p3.2 (which is a preview release well before 0.6).

By: Tink

Tink — Mon, 01 Jun 2009 22:07:35 +0000

Interesting question. I carried out the testing independently myself, without contacting the screen reader manufacturers.

Testing was done on three separate machines, using Charles Chen’s excellent test cases. Testing is rarely infallible, but the results appeared consistent over several iterations.

I’ll update the post to make this more apparent. If you’ve done any testing in this area yourself though, I’d be interested to hear of it.

By: Stephen Guerra

Stephen Guerra — Mon, 01 Jun 2009 21:39:49 +0000

I am suspect of the testing as its undetermined who is the tester and if the comments are just being offered by Tech Support by each manufacturer. I am further suspicious that the author of this blog in attempts to reach the different AT Manufacturers possibly did not get responses to the their queries causing by default a negative response. So truly how valid is this testing?