W3C home > Mailing lists > Public > public-pfwg@w3.org > February 2015

RE: Per discussion with the ARIA-APG group this morning, live region test files attached

From: Bryan Garaventa <bryan.garaventa@ssbbartgroup.com>
Date: Tue, 17 Feb 2015 18:30:27 +0000
To: Matthew King <mattking@us.ibm.com>
CC: "lwatson@paciellogroup.com" <lwatson@paciellogroup.com>, PF <public-pfwg@w3.org>
Message-ID: <BY2PR03MB347C978D21377F5C1E2F814982F0@BY2PR03MB347.namprd03.prod.outlook.com>
Thanks, those results aren't surprising. Basically all of these test files consist of one Test button with an onClick that triggers dynamic numerical content to be incremented to see if it is being announced properly.

The important thing to note here, is that all of the test files should be tested using the same AT to see how support differs for each implementation, which are all of the possible variations for Live Regions according to spec. If nothing or the wrong thing is announced, this is a bug. The Announce All folder includes sample files that should announce all of their content when new text is added, and the Announce New Text Only folder includes files that should only announce the newest added text and nothing else that the same container includes.

Where you note the word 'removed' isn't being announced in IE, actually the problem is greater than that, the removed content isn't being announced, but rather the newly added text, which is the opposite of what Firefox is doing by announcing only the removed text as the ARIA implementation is supposed to be doing in that case. So basically, in IE, the removal example is doing the same thing as aria-live='polite', and not what it should be doing by ignoring the newly added text and announcing only the text that is removed.

>From what I can see, live region support is only reliable in Firefox, and nowhere else, and this is using the leading desktop screen readers JAWS and NVDA. In fact, the level of support in IE is so bad that the majority of live region implementations are practically useless, and the disparity between IE and Firefox when using the same AT is so pronounced that it is actually sort of shocking.

There also seems to be a difference between the use of JAWS16 and IE10 versus IE11 on Win7 versus Win8, making me wonder if the correct events are even being fired by IE when live regions are implemented, which, if the case, goes beyond entering bugs with relevant screen reader venders.

From: Matthew King [mailto:mattking@us.ibm.com]
Sent: Monday, February 16, 2015 4:49 PM
To: Bryan Garaventa
Cc: lwatson@paciellogroup.com; PF
Subject: Re: Per discussion with the ARIA-APG group this morning, live region test files attached

Bryan, I tried a few of the files; have not had time to do all. Are there some that you think show defects and thus should be a priority to test?

I definitely found a few defects in the below.

I don't have IE 11; tried with IE 10, Firefox 35, JAWS 16, and NVDA 2014.4.

Note: in IE 10, I had to tab to the button and press space with vc off; the scripts didn't start when using only the virtual cursor. One time I had to use the JAWS cursor to get it to run.

Note: I added a few paragraphs of content to the files and read them to test politeness.

JAWS/IE/FF: announces changes and respects politeness.
NVDA/IE/FF: announces changes and respects politeness.

* JAWS/FF: Announces each removal and says "Removed " just before announcing the removed content. Respects politeness.
* JAWS/IE: JAWS announces the removed content, but it does not say "removed" before each removal in IE like it does in Firefox. Politeness is respected.
* NVDA does not announce removes in either Firefox or IE.

* JAWS/FF: announces all changes and respects politeness.
* JAWS/IE: The DOM updates are not recognized in the virtual buffer; must press insert+escape to see additions.
* NVEDA/FF: announces all changes and respects politeness.
* NVDA/IE: announces all changes and respects politeness. Unexpectedly, also announces changes when IE does not have focus.

Matt King
IBM Senior Technical Staff Member
I/T Chief Accessibility Strategist
IBM BT/CIO - Global Workforce and Web Process Enablement
Phone: (503) 578-2329, Tie line: 731-7398

From:        Bryan Garaventa <bryan.garaventa@ssbbartgroup.com<mailto:bryan.garaventa@ssbbartgroup.com>>
To:        PF <public-pfwg@w3.org<mailto:public-pfwg@w3.org>>,
Cc:        "lwatson@paciellogroup.com<mailto:lwatson@paciellogroup.com>" <lwatson@paciellogroup.com<mailto:lwatson@paciellogroup.com>>, Matthew King/Fishkill/IBM@IBMUS
Date:        02/16/2015 01:07 PM
Subject:        Per discussion with the ARIA-APG group this morning, live region test files attached

Attached is a zip file that includes the various live region test files that we spoke of earlier, which include all of the variations documented in the spec for the use of aria-live, role=status, role=log, and role=alert.

>From running my own tests, it appears that JAWS16 has broken support for most of these in IE11, though they have very good support in Firefox.

If you can confirm this as reproducible, I'll file an issue regarding this with FS including the same test files.

Somebody suggested that this issue may be the result of IE not firing the correct events when the live region is being written to, however I haven't found an accessible way of testing this as yet, so if anybody can confirm whether IE is in fact firing the correct events, please let me know.


[attachment "lr.zip" deleted by Matthew King/Fishkill/IBM]
Received on Tuesday, 17 February 2015 18:31:08 UTC

This archive was generated by hypermail 2.4.0 : Friday, 17 January 2020 16:45:23 UTC