E-mail List Archives

Search E-mail List Archives for

Number of posts in this thread: 6 (In chronological order)

From: Jim Homme
Date: Wed, Apr 13 2022 8:15AM
Subject: ARIA Polite VS Assertive
No previous message | Next message →

Hi,
I need to support a chat system and it needs to work for all screen readers. Does anyone know the differences between how JAWS handles ARIA assertive and polite messages VS how JAWS handles them? The same goes for Narrator, VoiceOver, and TalkBack. And are there differences among browser combinations?

Thanks for any and all help.

Jim

Thanks.

Jim


=========Jim Homme
Senior Digital Accessibility Consultant
Bender Consulting Services
412-787-8567
https://www.benderconsult.com/
Support the dreams of independence through employment for students with disabilities with your Amazon purchases.
https://smile.amazon.com/ch/83-0988251

From: glen walker
Date: Wed, Apr 13 2022 8:34AM
Subject: Re: ARIA Polite VS Assertive
← Previous message | Next message →

I rarely use assertive and use role=alert even less than that. Unless the
house is on fire, I just don't find that there are many messages that are
so important that the screen reader needs to be interrupted to announce an
update.

For example, in a chat interface, if I'm in a conversation and I'm typing a
response and listening to what I'm typing and then the other person sends a
response, I don't want my train of thought interrupted by hearing what they
typed in the middle of what I'm typing. I'm ok hearing their response when
I pause my typing. Now, one might argue that the other person's response
might affect what I'm typing. That's true but to me does not justify
interrupting what I was doing. But this is all just my personal preference.

You also have to consider that assertive regions might clear any pending
messages. The aria-live spec says, "User agents or assistive technologies
MAY choose to clear queued changes when an assertive change occurs." I'm
not sure which screen readers actually do that or if there are heuristics
or other factors that might cause it to happen. That is, JAWS or NVDA or
Voiceover or whatever might clear a pending polite announcement if an
assertive message came through in one situation but might not clear it in
another situation. To me it's too undefined on when that might happen and
goes back to whether messages are important enough to interrupt what is
happening.


On Wed, Apr 13, 2022 at 8:15 AM Jim Homme < = EMAIL ADDRESS REMOVED = > wrote:

> Hi,
> I need to support a chat system and it needs to work for all screen
> readers. Does anyone know the differences between how JAWS handles ARIA
> assertive and polite messages VS how JAWS handles them? The same goes for
> Narrator, VoiceOver, and TalkBack. And are there differences among browser
> combinations?
>
> Thanks for any and all help.
>
> Jim
>
> Thanks.
>
> Jim
>
>
> =========> Jim Homme
> Senior Digital Accessibility Consultant
> Bender Consulting Services
> 412-787-8567
> https://www.benderconsult.com/
> Support the dreams of independence through employment for students with
> disabilities with your Amazon purchases.
> https://smile.amazon.com/ch/83-0988251
>
> > > > >

From: Jim Homme
Date: Wed, Apr 13 2022 8:50AM
Subject: Re: ARIA Polite VS Assertive
← Previous message | Next message →

Hi Glen,
It sounds like you're saying that you prefer polite.

Thanks.

Jim

=========Jim Homme
Senior Digital Accessibility Consultant
Bender Consulting Services
412-787-8567
https://www.benderconsult.com/
Support the dreams of independence through employment for students with disabilities with your Amazon purchases.
https://smile.amazon.com/ch/83-0988251

-----Original Message-----
From: WebAIM-Forum < = EMAIL ADDRESS REMOVED = > On Behalf Of glen walker
Sent: Wednesday, April 13, 2022 10:34 AM
To: WebAIM Discussion List < = EMAIL ADDRESS REMOVED = >
Subject: Re: [WebAIM] ARIA Polite VS Assertive

I rarely use assertive and use role=alert even less than that. Unless the house is on fire, I just don't find that there are many messages that are so important that the screen reader needs to be interrupted to announce an update.

For example, in a chat interface, if I'm in a conversation and I'm typing a response and listening to what I'm typing and then the other person sends a response, I don't want my train of thought interrupted by hearing what they typed in the middle of what I'm typing. I'm ok hearing their response when I pause my typing. Now, one might argue that the other person's response might affect what I'm typing. That's true but to me does not justify interrupting what I was doing. But this is all just my personal preference.

You also have to consider that assertive regions might clear any pending messages. The aria-live spec says, "User agents or assistive technologies MAY choose to clear queued changes when an assertive change occurs." I'm not sure which screen readers actually do that or if there are heuristics or other factors that might cause it to happen. That is, JAWS or NVDA or Voiceover or whatever might clear a pending polite announcement if an assertive message came through in one situation but might not clear it in another situation. To me it's too undefined on when that might happen and goes back to whether messages are important enough to interrupt what is happening.


On Wed, Apr 13, 2022 at 8:15 AM Jim Homme < = EMAIL ADDRESS REMOVED = > wrote:

> Hi,
> I need to support a chat system and it needs to work for all screen
> readers. Does anyone know the differences between how JAWS handles
> ARIA assertive and polite messages VS how JAWS handles them? The same
> goes for Narrator, VoiceOver, and TalkBack. And are there differences
> among browser combinations?
>
> Thanks for any and all help.
>
> Jim
>
> Thanks.
>
> Jim
>
>
> =========> Jim Homme
> Senior Digital Accessibility Consultant Bender Consulting Services
> 412-787-8567
> https://www.benderconsult.com/
> Support the dreams of independence through employment for students
> with disabilities with your Amazon purchases.
> https://smile.amazon.com/ch/83-0988251
>
> > > archives at http://webaim.org/discussion/archives
> >

From: Mark Magennis
Date: Wed, Apr 13 2022 8:53AM
Subject: Re: ARIA Polite VS Assertive
← Previous message | Next message →

That's funny, I very rarely use polite. I guess it depends on how you judge the importance. But one issue with polite that you should be aware of is that it may queue messages and read them all at once when it gets a chance. This can be a disaster. For example, if you have a "remaining characters" count for a text input field. Putting the number in an assertive live region usually works quite well because it starts of with "remaining characters: 250" and as the user types, every now and then when they pause it says "243", "227", etc. But if you use aria-live="polite" it may wait until the user pauses and then blurt out "249 248 247 246 ...". Can be a pain. It depends on how fast the user types, how long they pause, etc. but usually an assertive region works best. Probably better still is scripting to only read out at intervals, say 250, 200, 150, 100, 50, 20, 10, 5, 1 but that's more dev effort.

So what's best often depends a lot on the nature of the task, e.g. chat vs. remaining character count

Mark

-----Original Message-----
From: WebAIM-Forum < = EMAIL ADDRESS REMOVED = > On Behalf Of glen walker
Sent: 13 April 2022 15:34
To: WebAIM Discussion List < = EMAIL ADDRESS REMOVED = >
Subject: [EXTERNAL] Re: [WebAIM] ARIA Polite VS Assertive

I rarely use assertive and use role=alert even less than that. Unless the house is on fire, I just don't find that there are many messages that are so important that the screen reader needs to be interrupted to announce an update.

For example, in a chat interface, if I'm in a conversation and I'm typing a response and listening to what I'm typing and then the other person sends a response, I don't want my train of thought interrupted by hearing what they typed in the middle of what I'm typing. I'm ok hearing their response when I pause my typing. Now, one might argue that the other person's response might affect what I'm typing. That's true but to me does not justify interrupting what I was doing. But this is all just my personal preference.

You also have to consider that assertive regions might clear any pending messages. The aria-live spec says, "User agents or assistive technologies MAY choose to clear queued changes when an assertive change occurs." I'm not sure which screen readers actually do that or if there are heuristics or other factors that might cause it to happen. That is, JAWS or NVDA or Voiceover or whatever might clear a pending polite announcement if an assertive message came through in one situation but might not clear it in another situation. To me it's too undefined on when that might happen and goes back to whether messages are important enough to interrupt what is happening.


On Wed, Apr 13, 2022 at 8:15 AM Jim Homme < = EMAIL ADDRESS REMOVED = > wrote:

> Hi,
> I need to support a chat system and it needs to work for all screen
> readers. Does anyone know the differences between how JAWS handles
> ARIA assertive and polite messages VS how JAWS handles them? The same
> goes for Narrator, VoiceOver, and TalkBack. And are there differences
> among browser combinations?
>
> Thanks for any and all help.
>
> Jim
>
> Thanks.
>
> Jim
>
>
> =========> Jim Homme
> Senior Digital Accessibility Consultant Bender Consulting Services
> 412-787-8567
> https://www.benderconsult.com/
> Support the dreams of independence through employment for students
> with disabilities with your Amazon purchases.
> https://smile.amazon.com/ch/83-0988251
>
> > > archives at http://webaim.org/discussion/archives
> >

From: glen walker
Date: Wed, Apr 13 2022 9:17AM
Subject: Re: ARIA Polite VS Assertive
← Previous message | Next message →

Yes, it certainly depends. On a "characters left" message, I never update
the live region on every character typed just like I don't update a
countdown timer on every second (assuming there is more than 10 seconds
left). It's really not much dev effort to get a nice effect. I usually
have two regions. One is visible and not live and is updated on every
keystroke or countdown timer second. The other is visually hidden
(sr-only) and is live (polite) and is updated periodically, such as after
every 50 keystrokes (if the characters left is a large number) or after
every minute (if there are several minutes left).

I then adjust the live updates as you get towards the end. With
keystrokes, if you're below 10 chars left, I'll update the live on every
keystroke. If there's a timer, then I'll update perhaps every 10 seconds
when it's under a minute left and maybe every second when there's 10
seconds left, or something like that. Is there a potential to hear
10,9,8,7... blasted at once? Maybe, but the advantages of everything else
working great outweigh the possibility of the one bad situation, in my
opinion.

But as I mentioned before, that's just my personal preference. There are
certainly other creative ways of handling this.

From: Jim Homme
Date: Wed, Apr 13 2022 9:38AM
Subject: Re: ARIA Polite VS Assertive
← Previous message | No next message

Hi,
Looking for a starting point to see the differences between how screen readers handle this, then possibly adjusting.

Thanks.

Jim

=========Jim Homme
Senior Digital Accessibility Consultant
Bender Consulting Services
412-787-8567
https://www.benderconsult.com/
Support the dreams of independence through employment for students with disabilities with your Amazon purchases.
https://smile.amazon.com/ch/83-0988251

-----Original Message-----
From: WebAIM-Forum < = EMAIL ADDRESS REMOVED = > On Behalf Of glen walker
Sent: Wednesday, April 13, 2022 11:17 AM
To: WebAIM Discussion List < = EMAIL ADDRESS REMOVED = >
Subject: Re: [WebAIM] ARIA Polite VS Assertive

Yes, it certainly depends. On a "characters left" message, I never update the live region on every character typed just like I don't update a countdown timer on every second (assuming there is more than 10 seconds left). It's really not much dev effort to get a nice effect. I usually have two regions. One is visible and not live and is updated on every keystroke or countdown timer second. The other is visually hidden
(sr-only) and is live (polite) and is updated periodically, such as after every 50 keystrokes (if the characters left is a large number) or after every minute (if there are several minutes left).

I then adjust the live updates as you get towards the end. With keystrokes, if you're below 10 chars left, I'll update the live on every keystroke. If there's a timer, then I'll update perhaps every 10 seconds when it's under a minute left and maybe every second when there's 10 seconds left, or something like that. Is there a potential to hear 10,9,8,7... blasted at once? Maybe, but the advantages of everything else working great outweigh the possibility of the one bad situation, in my opinion.

But as I mentioned before, that's just my personal preference. There are certainly other creative ways of handling this.