RE: role="list", Safari and VoiceOver

Hi again,

I've been doing some testing on a ul with/without list-style: none/role: list, using Browserstack (https://www.browserstack.com/),

- Mac OS Catalina + Safari 13.1 + VoiceOver

- UL list style: none -> VoiceOver doesn't anounce number of list items (that's the problem we are adressing in this thread).
- UL list style: none + role="list" -> VoiceOver does anounce number of list items (that's the solution we are using).
- Mac OS Big Sur + Safari 14.1 + VoiceOver

- UL list style: none -> VoiceOver does anounce number of list items.
- Mac OS Monterey + Safari 15.3 + VoiceOver

- UL list style: none -> VoiceOver does anounce number of list items.

So I think that, since Safari 14.1 it is not necessary to add role="list" to lists with list-style: none, because VoiceOver is announcing the number of list items.

I haven't seen any reaction to this fact, so I'd like somebody else to test it as I've done, in order to confirm that Apple has changed its mind and now Safari's behavior is like that of other browsers.

Thanks.

--
Bernat Lleonart

Enviat amb h [ProtonMail](https://protonmail.com/) correu segur.

------- Original Message -------
El dilluns, 2 de maig 2022 a les 14:43, Steve Green <steve.green@testpartners.co.uk> va escriure:

> Announcing “bullet” is useful when list items contain multiple paragraphs. Without it, you don’t know where one list item ends and the next one starts.
>
> Steve Green
>
> Managing Director
>
> Test Partners Ltd
>
> From: Taliesin Smith <talilief@gmail.com>
> Sent: 02 May 2022 15:22
> To: Marc Haunschild <marc.haunschild@accessibility.consulting>
> Cc: John Foliot <john@foliot.ca>; Peter Weil <peter.weil@wisc.edu>; Bernat Lleonart <bernat.lleonart@proton.me>; ML W3C, WAI <w3c-wai-ig@w3.org>
> Subject: Re: role="list", Safari and VoiceOver
>
> I agree, this is a very interesting discussion.
>
> I design interactive described experiences, and many screen reader users have said to me that they like knowing the number of items that are contained in a list and which item they are on. These useful details allow them to stop reading the list and simultaneously know how much they are skipping over.
>
> Headings, paragraphs, and lists, all provide very useful semantics…announcing “bullet” in addition to the list count seems annoyingly unnecessary, and would be nice if Apple respected this design standard.
>
> Taliesin Smith
>
> talilief@gmail.com
>
> ~.~.~
> Also reachable at:
>
> Taliesin.Smith@colorado.edu
> Inclusive Design Researcher
> PhET Interactive Simulations
> https://phet.colorado.edu/en/accessibility
> Physics Department
> University of Colorado, Boulder
>
>> On May 1, 2022, at 12:02 PM, Marc Haunschild <marc.haunschild@accessibility.consulting> wrote:
>>
>> Interesting discussion - I read, that users complain a lot about too many lists and that most users appreciate apples design decision.
>>
>> Sorry, but I don’t find this resource anymore. But maybe you have more reliable studies/ insights, that users want to have the lists back?
>>
>> Greetings
>>
>> mh
>>
>> --
>>
>> Mit freundlichen Grüßen
>>
>> Marc Haunschild
>> Accessibility Consultant
>>
>> [https://Accessibility.Consulting](https://accessibility.consulting/)
>>
>>> Am 29.04.2022 um 18:23 schrieb John Foliot <john@foliot.ca>:
>>>
>>> Hi  Bernat,
>>>
>>> A few thoughts: first, the arrogance of Apple is well documented, and this is not the only example of their tools not supporting web standards properly. The hubris of their decision policies is extremely frustrating.
>>>
>>> While I understand the desire to have "error-free" HTML, I think there are a few bullet points here to go back to your powers-that-be to discuss:
>>>
>>> a) due to a bug in Apple's product, their tool is not meeting the expected standard. Applying the fix, while not technically conformant, addresses the real need and gap that users experience.
>>>
>>> (Perhaps also remembering the design principle that drove the development of HTML 5 may help here: [the priority of constituents](https://www.w3.org/TR/html-design-principles/#priority-of-constituencies) then was summed up as: "In case of conflict, consider users over authors over implementors over specifiers over theoretical purity'".) Here, I would argue that knowing the rules, and the impact on breaking the rules, should be the final arbiter, and specifically adding that role - even though the validator is squawking - is the better choice for the accessible user-experience. Choose users over theoretical purity!
>>>
>>> b) It may also be worth remembering that back when ARIA was first introduced, the ARIA spec itself was 'vetted' and approved by the W3C, yet if/when you added ARIA attributes to a valid HTML 4.1 document back then, the validator complained that the resulting code was invalid. But as an informed developer, I knew that if my HTML 4.1 code was valid without the ARIA attributes, that adding those attributes really wasn't creating harm, and in fact was making things EVEN BETTER.
>>>
>>> So, really, the agency where you work sort of need to choose between two objectives: error-free HTML, or optimized accessibility for all users on all platforms/user-agents. (For me it's a no-contest decision, but I appreciate that it may not be that way everywhere).
>>>
>>> HTH
>>>
>>> JF
>>>
>>> On Fri, Apr 29, 2022 at 11:33 AM Peter Weil <peter.weil@wisc.edu> wrote:
>>>
>>>> Bernat,
>>>>
>>>> For what it’s worth, I use role=“list” in these situations, and respectfully disagree with the reasons given by the Safari team to change the semantics. See https://developer.mozilla.org/en-US/docs/Web/CSS/list-style for a more complete explanation. I would ignore the validator warning. If you want the list semantics to be there in Safari, then it *is* necessary.
>>>>
>>>> I’ll let someone else answer the other question.
>>>>
>>>> Peter
>>>>
>>>>> On Apr 27, 2022, at 3:43 AM, Bernat Lleonart <bernat.lleonart@proton.me> wrote:
>>>>>
>>>>> Hi,
>>>>>
>>>>> Some time ago I came across this blog post: https://www.scottohara.me/blog/2019/01/12/lists-and-safari.html and this tweet: https://twitter.com/rogerjohansson/status/1083600157076324352
>>>>>
>>>>> Since then I'm adding role="list" to the lists that have list-style: none applied.
>>>>>
>>>>> However, the w3c validator show a warning: "The list role is unnecessary for element ul."
>>>>>
>>>>> In the web design agency I'm working I'm asked to produce error/warning free html code.
>>>>>
>>>>> So my questions are:
>>>>>
>>>>> • Which CSS properties/values remove list semantics apart from list-style: none? For example: does display: flex remove list semantics?
>>>>> • What's your approach around this issue? Do you use role="list" on lists with list-style: none?
>>>>> • Since O'Hara's post is from 2019, I wonder if VoiceOver/Safari approach is still the same.
>>>>>
>>>>> Thanks.
>>>>>
>>>>> --
>>>>> Bernat Lleonart
>>>>>
>>>>> Enviat amb h ProtonMail correu segur.
>>>>
>>>> --
>>>> Peter Weil
>>>> Web Developer
>>>> University Marketing, University of Wisconsin–Madison
>>>> 608-220-3089
>>>
>>> --
>>>
>>> John Foliot | Senior Industry Specialist, Digital Accessibility |
>>> W3C Accessibility Standards Contributor |
>>>
>>> "I made this so long because I did not have time to make it shorter." - Pascal "links go places, buttons do things"