Re: Query regarding 1.3.4

Brooks,

I was going to let this run its course, but I must take exception to some
of the assertions you are making:

> I completely disagree with the notion that accepting a refresh press in
the browser (also known as abandoning the process) is a fix-all for
non-compliant content.

*It's not*, It is a single Technique for meeting the *requirements of a
single Success Criteria. *

Accessibility and accessibility conformance is subtle, it's not a brute
force hammer. And as you and others have noted, an
accessible user-experience is not a single SC, it's the sum of all of them
put together. When you read the Understanding Document associated to SC
1.3.4, you will note that it states, "*The goal of this Success Criterion
is that authors should never restrict content's orientation, thus ensuring
that it always match the device display orientation.*" That's it, just as
the goal of SC 2.4.2 (Page Titled) is "*...to help users find content and
orient themselves within it by ensuring that each Web page has a
descriptive title.*"

But, (I hear some ask), what does that have to do with page orientation?
Nothing - but that's the point. You could pass SC 1.3.4 in all of the
permutations you envision, but if you fail to provide a page <title> you've
STILL got an "inaccessible experience" for some users. And a page refresh
in that use-case is not a fix-all for that problem.


> I think it is unreasonable to allow a page refresh to serve as a pass for
orientation content/functionality failures without clear instructions at
the beginning of the page to do so.

That is not an unreasonable opinion, however individual opinion does not a
Success Criteria make. I think that in principle what you are stating is
indeed a Best Practice, but I also have to wonder aloud if this isn't a
tempest in a teapot - do you have any real evidence of users, who, *midway
through filling out a form*, decide to change orientation of that form? I'm
not suggesting that it would NEVER happen, but the law of averages and
logic & reason strikes me that your strawman scenario is quite the edge
case: that most users will have "oriented" their view long before they dive
into a form.


> Seems like a lot of excuse making to give an inaccessible experience an
easy pass.

Again, I believe you are conflating this into far more than it is. NOBODY
has said that the technique of invoking a page refresh in this scenario was
an awesome user experience - it isn't. Nor is anyone saying that it is a
"Pass" on an "inaccessible experience" - because it isn't - but it IS a
successful meeting of the mandated outcome of SC 1.3.4.

The goal here (going all the way back to Ramakrishnan's original
question) is to *evaluate a single Success Criteria*, one of many that add
up to creating a MORE accessible experience for MOST users. And the hard
reality is, to meet the single goal of rendering content in a viewport that
has experienced a physical change in orientation is that, under some
conditions, some content may need to be refreshed to render properly -
although some quick testing here confirms to me that most forms, even when
you change orientation mid-way through completing them, will a) adapt to
the new screen parameters, and b) retain all of the content previously
entered.

So to arrive at the scenario where a hard page-refresh is required to
ensure a partially completed form is fully visible and "actionable" after a
change of orientation means that there is likely a lot more funkyness on
that page - there will also likely be other SC failures (I cannot for one
second believe that a web page would pass all of the other 49 A & AA SC in
WCAG 2.1 and ONLY experience this one failure - I just can't).

For me, Accessibility has always been about removing barriers for users,
and/but when the barrier cannot be fully removed, can it be minimized? (I
learned this the hard way back when I worked at Chase Bank - where US
banking regulations and serious and legitimate security concerns over DOS
attacks - based on actual evidence - trumped eliminating *ALL* CAPTCHAS
from their sites. We got rid of a lot of them, but had to accept some water
with our wine in some instances, our objections notwithstanding. Nobody on
the Accessibility team back then was happy about that, but we had no choice
but to accept the trade-off: a major bank site taken offline by a DOS
attack is inaccessible to everyone!)

And so personally I've landed at understanding that an all-or-nothing
stance may feel righteous, but at the end of the day getting perfect is (by
my experience) unachievable - there are always trade offs. Accessibility
is, and always will be, a long-tail problem in part because we have
'solutions' and articulated needs authored 'at scale' - based on the
understood needs of 'groups of users': blind/low vision, mobility impaired,
cognitively impaired, etc. Yet because each person is unique, so too will
be their needs (this is also why meeting the needs of ALL users with
cognition impairments is so complex - the different types of issues along
with the multiple degrees of severity and impact).

The sad reality today is that even if a website successfully met all 50 A
and AA SC in WCAG 2.1, that the content may still be inaccessible to some
users, based on that user's individual needs. Perfection is an ideal, not
an outcome.

With respect,

JF

On Wed, Nov 9, 2022 at 6:50 PM Brooks Newton <brooksallennewton@gmail.com>
wrote:

> Hi,
>
> All fair points, although I completely disagree with the notion that
> accepting a refresh press in the browser (also known as abandoning the
> process) is a fix-all for non-compliant content.  I do appreciate the clear
> explanation from both of you, Juliette and John, as to your perspectives.
> I was on all of those calls when this was discussed, and likely scribed for
> a session or two during meetings on this very topic. That's the good thing
> about being in the room when something happens - you've have first-hand
> knowledge of what could never be believed second-hand.
>
> I never had any expectation that this normative definition would exclude
> the scenario that I've laid out and that others have said happens more
> often than you'd expect.
>
> Long story short from me: I think it is unreasonable to allow a page
> refresh to serve as a pass for orientation content/functionality failures
> without clear instructions at the beginning of the page to do so.  If it's
> a known issue, content authors have an obligation to inform users to
> perform a refresh before beginning to fill out the form.
>
> Seems like a lot of excuse making to give an inaccessible experience an
> easy pass.  That happens a lot, especially with browser bugs or innately
> inaccessible practices that are controlled by software.
>
> Brooks
>
> On Wed, Nov 9, 2022 at 3:56 PM John Foliot <john@foliot.ca> wrote:
>
>> Hi Brooks,
>>
>> >  Is anyone in this thread saying that abandoning the user task and
>> refreshing the page is a viable reason for passing SC 1.3.4 - Orientation
>> <https://www.w3.org/WAI/WCAG21/Understanding/orientation.html>  in the
>> half-finished business license application scenario I presented?  If so,
>> really?
>>
>>
>> Viable reason? I don't understand that - what do you mean by "reason"?
>>
>> Are you asking, does the current technique -  G214
>> <https://www.w3.org/WAI/WCAG21/Techniques/general/G214> "Using a control
>> to allow access to content in different orientations which is otherwise
>> restricted" suffice in meeting this single Success Criteria? Yes - really.
>> It may not solve ALL problems (and may potentially introduce others), but
>> it DOES address this problem.
>>
>> Does it absolve content authors from also having to ensure other aspects
>> of their content is accessible?
>>
>> No. But we cannot try and stretch a SC to meet all possible scenarios,
>> especially when there is a confluence of multiple potential issues. As AWK
>> noted, it is less than ideal, but IF the content needs to be refreshed, BUT
>> (and?) the content author is also supporting Success Criterion 3.3.9, then
>> why would you fail this? And if using that technique DID erase previous
>> user entered data, well, that's a failure of SC 3.3.9, and NOT of SC 1.3.4.
>>
>> > Maybe what's missing here is the recognition that multiple other
>> success criteria will be failed in this scenario, so we don't have to pin
>> the failure on SC 1.3.4?
>>
>> Precisely!
>> WCAG is like a jigsaw puzzle, all of the SC are required to complete the
>> puzzle, and there is often an inter-dependency of multiple SC contributing
>> to the accessible experience. For example, to make a video truly accessible
>> usually requires Captions (SC 1.2.2) AND Audio Description (SC 1.2.3 and/or
>> SC 1.2.5), and I often argue that for some users (deaf/blind) a full
>> transcript should also be provided (SC 1.2.3 OR SC 1.2.8 AAA).
>>
>> A page's content could easily Pass SC 1.2.2 while also failing SC 1.2.5,
>> and we see this frequently, and yet this rarely comes up as "confusing".
>> Yet intuitively, we all know that the video isn't "fully" accessible, but
>> it is partially so. How is that any different than your use-case?
>>
>> I personally think that a large part of the problem here is that too many
>> folks dive into Techniques, without stopping and truly thinking about the
>> normative text of the SC, which for 1.3.4 is pretty simple: content has to
>> "work" (render, interact, etc.) in either view format, as decided upon by
>> the end user.
>>
>> If content authors deliberately lock content into a fixed view, then the
>> requirement is for the end-user to be able to override that - the old W3C
>> axiom "author proposes, user disposes". Depending on *how *the author
>> "locked down" the view however, it may be as simple as invoking a
>> page-refresh, or it may be more complicated, with a dedicated "unlock"
>> button/control also furnished by the content author... it all depends on
>> how the locking is achieved.
>>
>> I'll also question your use-case (half-finished business license
>> application) as being fairly niche - I'm not saying that some users might
>> re-orient their web form after completing half of a form, but that does
>> strike me as odd... I mean, I could also change the font-face or color or
>> size midway through a form as well, but (to use your expression), Really?
>>
>> Nonetheless, if the user does that, and is obligated to do a screen
>> refresh to get it to render correctly, then it is passing this single SC -
>> nothing more, nothing less. Nobody is suggesting that this is optimal, but
>> a strict reading of the SC confirms that at least this one specific SC has
>> been met. Bravo, 1 down, another 49 to go...
>>
>> Peace out friend
>>
>> JF
>>
>> On Wed, Nov 9, 2022 at 3:24 PM Brooks Newton <brooksallennewton@gmail.com>
>> wrote:
>>
>>> OK, so let me get this straight.  Is anyone in this thread saying that
>>> abandoning the user task and refreshing the page is a viable reason for
>>> passing SC 1.3.4 - Orientation
>>> <https://www.w3.org/WAI/WCAG21/Understanding/orientation.html>  in the
>>> half-finished business license application scenario I presented?  If so,
>>> really?
>>>
>>> How many success criteria do we have in WCAG 2.x that exist primarily to
>>> prevent users with disabilities from abandoning tasks that may be
>>> technically possible, but not probable due to user disabilities?   I know
>>> I'm asking a lot of questions.  This interpretation of SC 1.3.4  having an
>>> abandon/refresh exception doesn't seem in keeping with the spirit of many
>>> other success criteria in WCAG 2.x.
>>>
>>> Maybe what's missing here is the recognition that multiple other success
>>> criteria will be failed in this scenario, so we don't have to pin the
>>> failure on SC 1.3.4?  Like Alan, I'm having a difficult time justifying why
>>> the user scenario I wrote earlier in the thread wouldn't be a hard fail of
>>> SC 1.3.4, even if a refresh of the page fixed the content or functionality
>>> in the given display orientation.
>>>
>>> Are instructions provided to the user to simply abandon their task and
>>> refresh that page to get the page working right?  If they aren't provided,
>>> is that a failure of SC 3.3.2 - Labels or Instructions
>>> <https://www.w3.org/WAI/WCAG21/Understanding/labels-or-instructions.html>?
>>> Is it reasonable to expect that a user will just intuitively know that
>>> refreshing the page is the fix to the problem?  Can you expect the user to
>>> make that destructive choice when they've put a lot of time and energy into
>>> what they've already input or interacted with on the page?
>>>
>>> Isn't the disproportionate time and energy  required to complete tasks
>>> for people with certain disabilities the reason for supporting successful
>>> outcomes that are probable, not just theoretically possible? How long
>>> should it take for the user to reach that conclusion to throw away page
>>> progress by refreshing if they are never explicitly directed to do so by
>>> the content author?
>>>
>>> Would this orientation scenario also be a failure of SC 3.3.9 -
>>> Redundant Entry (WCAG 2.2)
>>> <https://www.w3.org/TR/WCAG22/#redundant-entry>, in the case of a
>>> half-filled out application that must be erased and filled out again after
>>> refresh?
>>>
>>> I'm not saying I have all of the answers here - I'm just trying to ask
>>> the right questions to get closer to consensus.
>>>
>>> Brooks Newton
>>>
>>> On Wed, Nov 9, 2022 at 12:49 PM Bristow, Alan <Alan.Bristow@elections.ca>
>>> wrote:
>>>
>>>> Thank you Giacomo​.
>>>>
>>>>
>>>> Regards,
>>>>
>>>> Alan
>>>> . . . . -   . . - - -
>>>> Alan Bristow ( he / him / il )
>>>> Web Developer / Développeur Web
>>>> Elections Canada / Élections Canada
>>>> alan.bristow@elections.ca
>>>> ------------------------------
>>>> *From:* Giacomo Petri <giacomopetri89@gmail.com>
>>>> *Sent:* Wednesday, November 9, 2022 1:07 PM
>>>> *To:* Marc Haunschild (Accessibility Consulting)
>>>> *Cc:* John Foliot; Kevin Prince; Ramakrishnan Subramanian;
>>>> w3c-wai-ig@w3.org
>>>> *Subject:* Re: Query regarding 1.3.4
>>>>
>>>> Ce message a été envoyé par un expéditeur externe. Veuillez faire
>>>> preuve de prudence et ne pas cliquer sur les liens ou ouvrir les pièces
>>>> jointes à moins de reconnaître l'expéditeur et de savoir que le contenu est
>>>> sûr.
>>>>
>>>> This message was sent from an external sender. Please exercise caution
>>>> and do not click links or open attachments unless you recognize the sender
>>>> and know the content is safe.
>>>>
>>>>
>>>> I assumed per 1.3.4 that both orientations work on page load but not
>>>> while switching from portrait to landscape and vice versa.
>>>> Especially after reading
>>>>
>>>> While the content may work perfectly in both and only fail when
>>>>> transitioning from one to the other, I see nothing that indicates working
>>>>> in the initial orientation is all that is needed. By definition orientation
>>>>> is subject to change, and so it seems pretty clear the intention is content
>>>>> works in either orientation, including after a change in orientation.
>>>>>
>>>>
>>>> from Alan.
>>>>
>>>> This case happens more frequently than expected, especially for
>>>> animated elements such as sliders or carousels.
>>>>
>>>> That's the reason of my feedback and why I've opened
>>>> https://github.com/w3c/wcag/issues/2771
>>>>
>>>>>
>>
>> --
>> *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"
>>
>

-- 
*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"