W3C home > Mailing lists > Public > public-silver@w3.org > March 2019

Re: WCAG 2.2 acceptance criteria

From: John Foliot <john.foliot@deque.com>
Date: Thu, 7 Mar 2019 16:08:35 -0600
Message-ID: <CAKdCpxwsLxvdV7+8U-tE7kwkkgs85VYT2n-Wpa9KGSSL-9DHMg@mail.gmail.com>
To: Chuck Adams <charles.adams@oracle.com>
Cc: Glenda Sims <glenda.sims@deque.com>, "Delisi, Jennie (MNIT)" <jennie.delisi@state.mn.us>, Alastair Campbell <acampbell@nomensa.com>, "lisa.seeman" <lisa.seeman@zoho.com>, COGA TF <public-cognitive-a11y-tf@w3.org>, Silver TF <public-silver@w3.org>, Andrew Kirkpatrick <akirkpat@adobe.com>
So... I *do not* want to be taken out of context here, but a few comments:

One example we discussed was the current testing required to ensure that
the appropriate alt text is assigned for each image used on a page. 1-2
images on a page, not a big deal to test.


This is actually more nuanced, because the Success Criteria does not call
for "appropriate", it demands "equivalent purpose" which may not always be
the same. <ing src="" alt="sunset"> would pass the SC as written, even
though most of us instinctively know that the alt text there is "weak" (and
further, some might argue, warranting a longer textual description). *But
it meets the legal requirement.* Facebook's AI will often provide an alt
text "may include two people in front of a car" [sic], which again isn't
great alt text, but it *does* meet the minimum bar. (And while I never
advocate for just the minimum bar, I am pragmatic enough to realize that
sometimes that's the best you're gonna get.)


But, on a catalogue page, it could be significant.


Maybe, maybe not. I could also envision a code block for that catalog page
that looked something like this: <img src="" alt="Photo: %item_name%">
<h3>%item_name%</h3>, where the value of %item_name% would, in both
instances, be populated from a data-base. Again, I'm not *advocating* for
that, I'm suggesting it is a reasonable solution in some scenarios, and
there the test could be as simple as looking at the source code external to
the data-base (or manually checking 3 images to ensure the pattern is
consistent, and then moving on)


The question came down to the concept that there may be manual testing that
(at this time) may be the only way to truly ensure a barrier is not met by
individuals with cognitive disabilities.


Sure, but as the requirements become more sophisticated, a specific testing
methodology must also be articulated and put into place: we can't just toss
requirements over the wall and hope everyone will figure it out on their
own. The Silver TF have discussed this at some length already (and AFAIK
not yet come up with a definitive "solution").

From a matter of equality standpoint, why would the testing to address the
needs for one group be ok if it takes a lot of time, because they got in on
the creation of success criteria at the beginning of the process; but for
another group who’s needs were addressed more thoroughly later in the
development of success criteria, manual testing that may sometimes require
some time cannot be considered?

Respectfully, I find that something of an antagonistic statement: this is
not singling out one group over another, it's about ensuring that what we
demand of content creators can be accurately and consistently verified for
compliance requirements. I would strongly caution that this discussion not
dive into one that pits one user group against the other: we're all here
for the same reasons. [*Flagging chairs*]


Meanwhile, Glenda wrote:

To something like this:

   - Be feasibly testable in a "reasonable amount of time" through
   automated or manual processes prior to Candidate Recommendation stage.
   Examples include:
      - Automated - an automated testing tool exists that quickly and
      accurately determines if the criteria is met or not.
      - Assisted - a software tool exists that makes it more efficient for
      a tester to accurately determines if the criteria is met or not.
      - Manual - a manual process exists that makes it possible for a
      tester to accurately determines if the criteria is met or not.

note:  "reasonable amount of time" can be determined by a call for
consensus.

I'd actually leave the time element on the cutting-room floor: a)
personally I don't think we'd ever find that magic number, and b) I vaguely
recall a SC that speaks about "Timing Adjustable" <grin>, which to
paraphrase, effectively states that we shouldn't be locking people into
specific time-frames, that they can "adjust" that timing to meet their
individual needs. I would think that this would be of particular interest
to the Coga TF, as I suspect this is a real issue for many of those in that
user-group.

I think what is far more important (I'd go as far as "Critical") is that we
produce, in conjunction with any SC that requires manual testing, a
specific testing methodology - a 'script' as it were - on how to
consistently test a component on the page, with clear and unambiguous
'markers' on what is sufficient versus what is not. It's the methodology
piece that is critical, not the time it takes to do it (for example, the
*only* way to accurately determine if Audio Description is "correct"
today is to watch the entire video with Audio Description turned on -
whether that video is 3 minutes or 30 minutes or 300 minutes...)

Minus the time reference however, +1 to Glenda's suggestion.

JF


On Thu, Mar 7, 2019 at 3:31 PM Chuck Adams <charles.adams@oracle.com> wrote:

> +1 Chuck
>
>
>
> *From:* Glenda Sims <glenda.sims@deque.com>
> *Sent:* Thursday, March 7, 2019 2:03 PM
> *To:* Delisi, Jennie (MNIT) <jennie.delisi@state.mn.us>
> *Cc:* John Foliot <john.foliot@deque.com>; Alastair Campbell <
> acampbell@nomensa.com>; lisa.seeman@zoho.com; COGA TF <
> public-cognitive-a11y-tf@w3.org>; Silver TF <public-silver@w3.org>
> *Subject:* Re: WCAG 2.2 acceptance criteria
>
>
>
> Goodwitch magically appears after being MIA for weeks to say:
>
>
>
> I suggest we clarify this bullet a bit more.  I think the example is a
> useful example, but it isn't the only way to be "feasibly testable".  And
> the way the sentence is written, it is hard to parse/process.  So what if
> we changed from this:
>
>    - Be feasibly testable through automated or manual processes, i.e.
>    take a few minutes per page with tools available prior to Candidate
>    Recommendation stage.
>
> To something like this:
>
>    - Be feasibly testable in a "reasonable amount of time" through
>    automated or manual processes prior to Candidate Recommendation stage.
>    Examples include:
>
>
>    - Automated - an automated testing tool exists that quickly and
>       accurately determines if the criteria is met or not.
>       - Assisted - a software tool exists that makes it more efficient
>       for a tester to accurately determines if the criteria is met or
>       not.
>       - Manual - a manual process exists that makes it possible for a
>       tester to accurately determines if the criteria is met or not.
>
> note:  "reasonable amount of time" can be determined by a call for
> consensus.
>
>
>
> I'd suggest that if we pursue this "reasonable amount of time"
> angle...that it be based on "reasonable amount of time" to test an ELEMENT
> (not a page).  I think the variance in amount of time to test a page (when
> pages can endlessly scroll) will make it impossible to come up with a
> "reasonable amount of time" per page.
>
>
>
> I'm not in favor of leaving the requirement as it is currently drafted at
> https://www.w3.org/WAI/GL/wiki/WCAG_2.2_Success_criterion_acceptance_requirements
> <https://urldefense.proofpoint.com/v2/url?u=https-3A__www.w3.org_WAI_GL_wiki_WCAG-5F2.2-5FSuccess-5Fcriterion-5Facceptance-5Frequirements&d=DwMFaQ&c=RoP1YumCXCgaWHvlZYR8PZh8Bv7qIrMUB65eapI_JnE&r=b-9TIC95K-nLEKIDibNXAN_FKV-iXhLlAW2Zc3ebV_c&m=_p7Yxrl6Zp2sk72o5dwgmlwhjQ4eLo4fUWPVLEoAbk0&s=5ef2UFktuiL-eTBWN_T3f0qUokZodp6f36XFMELbhbU&e=>
>
>
>
>
> G
>
>
>
> *glenda sims* <glenda.sims@deque.com>, cpacc
> <https://urldefense.proofpoint.com/v2/url?u=http-3A__www.accessibilityassociation.org_certification&d=DwMFaQ&c=RoP1YumCXCgaWHvlZYR8PZh8Bv7qIrMUB65eapI_JnE&r=b-9TIC95K-nLEKIDibNXAN_FKV-iXhLlAW2Zc3ebV_c&m=_p7Yxrl6Zp2sk72o5dwgmlwhjQ4eLo4fUWPVLEoAbk0&s=9o5lmzpfCNSt1f4YcOrQ9SYxuekoXeU2JQrstNBegME&e=>
>  | team a11y lead | 512.963.3773
>
>
>
>         deque systems
> <https://urldefense.proofpoint.com/v2/url?u=http-3A__www.deque.com&d=DwMFaQ&c=RoP1YumCXCgaWHvlZYR8PZh8Bv7qIrMUB65eapI_JnE&r=b-9TIC95K-nLEKIDibNXAN_FKV-iXhLlAW2Zc3ebV_c&m=_p7Yxrl6Zp2sk72o5dwgmlwhjQ4eLo4fUWPVLEoAbk0&s=AOB6VeDCie4SvizdBULO7MzT1sYorNafNsRlwX7_YEo&e=>
>   accessibility for good
>
>
>
>
>
> On Thu, Mar 7, 2019 at 1:31 PM Delisi, Jennie (MNIT) <
> jennie.delisi@state.mn.us> wrote:
>
> Hello,
>
> Part of the concerns the COGA group discussed was that manual tests are
> often required, and the variety of time required to test different pages
> can vary greatly, depending on the content of that page.
>
>
>
> One example we discussed was the current testing required to ensure that
> the appropriate alt text is assigned for each image used on a page. 1-2
> images on a page, not a big deal to test. But, on a catalogue page, it
> could be significant.
>
> The question came down to the concept that there may be manual testing
> that (at this time) may be the only way to truly ensure a barrier is not
> met by individuals with cognitive disabilities.
>
>
>
> I, too, work in an environment where a lot of testing occurs every day.
> And, we have to hold contractors, vendors, and employees to standards that
> can be measured. We need to be able to provide detailed and consistent
> feedback when a failure of a success criteria has been noted. The time
> taken to complete testing is definitely important. But, consideration of
> barriers is the whole goal, right?
>
>
>
> From a matter of equality standpoint, why would the testing to address the
> needs for one group be ok if it takes a lot of time, because they got in on
> the creation of success criteria at the beginning of the process; but for
> another group who’s needs were addressed more thoroughly later in the
> development of success criteria, manual testing that may sometimes require
> some time cannot be considered?
>
>
>
> I would like to propose that the language about the time it takes to
> complete a test have an exception process, or propose a rewording of the
> time component, so that the barriers experienced by this group of
> individuals with disabilities receives fair consideration in this process.
>
>
>
> Jennie
>
>
>
> *Jennie Delisi, MA, CPWA*
>
> Accessibility Analyst | Office of Accessibility
>
> *Minnesota IT Services* |* Partners in Performance*
>
> 658 Cedar Street
>
> St. Paul, MN 55155
>
> O: 651-201-1135
>
> *Information Technology for Minnesota Government* | mn.gov/mnit
> <https://urldefense.proofpoint.com/v2/url?u=http-3A__mn.gov_mnit&d=DwMFaQ&c=RoP1YumCXCgaWHvlZYR8PZh8Bv7qIrMUB65eapI_JnE&r=b-9TIC95K-nLEKIDibNXAN_FKV-iXhLlAW2Zc3ebV_c&m=_p7Yxrl6Zp2sk72o5dwgmlwhjQ4eLo4fUWPVLEoAbk0&s=_LFhqvlrxsO3lXWJLlfp_27jFLhINYqTNbloiIO7Fuk&e=>
>
> [image: Minnesota IT Services Logo]
>
> [image: Facebook logo]
> <https://urldefense.proofpoint.com/v2/url?u=https-3A__www.facebook.com_MN.ITServices&d=DwMFaQ&c=RoP1YumCXCgaWHvlZYR8PZh8Bv7qIrMUB65eapI_JnE&r=b-9TIC95K-nLEKIDibNXAN_FKV-iXhLlAW2Zc3ebV_c&m=_p7Yxrl6Zp2sk72o5dwgmlwhjQ4eLo4fUWPVLEoAbk0&s=YvE-iUe2YSCmKu-16UikG2dQBqGSAzbXuMtLtuLpvt8&e=>[image:
> LinkedIn logo]
> <https://urldefense.proofpoint.com/v2/url?u=https-3A__www.linkedin.com_company_mn-2Dit-2Dservices&d=DwMFaQ&c=RoP1YumCXCgaWHvlZYR8PZh8Bv7qIrMUB65eapI_JnE&r=b-9TIC95K-nLEKIDibNXAN_FKV-iXhLlAW2Zc3ebV_c&m=_p7Yxrl6Zp2sk72o5dwgmlwhjQ4eLo4fUWPVLEoAbk0&s=RK5dztyDuf9FrOqeASJlnJCg223LbPm8cHpkQHow7co&e=>[image:
> Twitter logo]
> <https://urldefense.proofpoint.com/v2/url?u=https-3A__twitter.com_mnit-5Fservices&d=DwMFaQ&c=RoP1YumCXCgaWHvlZYR8PZh8Bv7qIrMUB65eapI_JnE&r=b-9TIC95K-nLEKIDibNXAN_FKV-iXhLlAW2Zc3ebV_c&m=_p7Yxrl6Zp2sk72o5dwgmlwhjQ4eLo4fUWPVLEoAbk0&s=rcsHJ6yQ3mOK0YLis1zBI7MepoVIPPbUJP_8TCpirLs&e=>
>
>
>
>
>
>
>
> *From:* John Foliot <john.foliot@deque.com>
> *Sent:* Thursday, March 7, 2019 11:26 AM
> *To:* Alastair Campbell <acampbell@nomensa.com>
> *Cc:* lisa.seeman@zoho.com; Delisi, Jennie (MNIT) <
> jennie.delisi@state.mn.us>; COGA TF <public-cognitive-a11y-tf@w3.org>;
> Silver TF <public-silver@w3.org>
> *Subject:* Re: WCAG 2.2 acceptance criteria
>
>
>
> Hi All,
>
>
>
> To perhaps also put a finer distinction on it... W3C Process mandates two
> independent implementations of whatever new technology is being proposed -
> a testing activity we actually did last spring during CSUN for the 2.1
> Success Criteria (where, for SC 1.3.6 @ AAA we actually used the
> implementations that Lisa had pointed us to). Those implementations may or
> may not also serve as a 'testing tool', but as the Silver discussion
> continues, a repeatable testing methodology will need to surface for each
> new requirement, whether that is via a tool (mechanical tests - see: ACT
> TF), or via a 'cognitive walk-though' or similar methodology (a process
> still to be fully defined in Silver).
>
>
>
> At the end of the day, while it is true that our primary audience is and
> will always be users with disabilities (of all stripes and forms), a second
> important consideration is compliance requirements mandated by legislation.
> To clear that hurdle, we will need to ensure that both implementers and
> consumers have a baseline measurable & impartial (non-subjective) "test",
> so that entities can then claim conformance based upon the outcome of said
> test.
>
>
>
> JF
>
>
>
> On Thu, Mar 7, 2019 at 10:52 AM Alastair Campbell <acampbell@nomensa.com>
> wrote:
>
> Hi Lisa,
>
>
>
> > To meet new user needs we may need new tools and reviews may need to
> acquire new skills and knowledge.
>
>
>
> Which is fine, perhaps we can clarify that it means available at the time
> of publication?
>
>
>
> New tools, especially if they “take a day” from a programmer would need to
> be available at the time of publication, for the reasons I outlined in the
> last email.
>
>
>
>
>
> > Also new tools will come as soon as we know a SC will be accepted. in
> other word at CR. With WCAGs current history it will not come before then.
>
>
>
> Can you point to a previous example? I.e. where a tool that didn’t exist
> was required to meet an SC wasn’t available until after CR?
>
> The closest I can think of is ARIA in WCAG 2.0, but it wasn’t actually
> required to meet the SCs.
>
>
>
> It is very difficult to deal something in CR which then has to be pulled
> because no one has created a tool, the whole timeline goes back a step. The
> way the W3C prefers to work is to have working prototypes/code created
> prior to specs. This has been a hard-learned approach [1].
>
>
>
> I suggest that if an SC needs a tool, we work up the SC template and go
> through the initial process. That could be accepted on the condition that a
> tool will be available. If it does not become available then the SC will be
> removed before CR.
>
>
>
> It would also help to put those SC(s) first so people have more time to
> work on the tools, I’ll make a note of that.
>
>
>
> Cheers,
>
>
>
> -Alastair
>
>
>
>
>
> 1] Accessibility example for what should be a ‘simple’ thing, the naming
> algorithm.
>
>
> https://www.linkedin.com/pulse/future-accname-spec-planning-strategy-functional-using-garaventa/
> <https://urldefense.proofpoint.com/v2/url?u=https-3A__gcc01.safelinks.protection.outlook.com_-3Furl-3Dhttps-253A-252F-252Fwww.linkedin.com-252Fpulse-252Ffuture-2Daccname-2Dspec-2Dplanning-2Dstrategy-2Dfunctional-2Dusing-2Dgaraventa-252F-26data-3D02-257C01-257Cjennie.delisi-2540state.mn.us-257C2a94ca2523bb46a0bdd208d6a321fd94-257Ceb14b04624c445198f26b89c2159828c-257C0-257C0-257C636875763714627907-26sdata-3DVCYgFIjR5CMjFxunizRlfRp8QYNbGpWZR8Sb6OmhcQI-253D-26reserved-3D0&d=DwMFaQ&c=RoP1YumCXCgaWHvlZYR8PZh8Bv7qIrMUB65eapI_JnE&r=b-9TIC95K-nLEKIDibNXAN_FKV-iXhLlAW2Zc3ebV_c&m=_p7Yxrl6Zp2sk72o5dwgmlwhjQ4eLo4fUWPVLEoAbk0&s=fQvWalvG5VAStfZ7Pmso1gyaTqOJ_sivl3M1isFCcBU&e=>
>
>
>
>
> --
>
> *​**John Foliot* | Principal Accessibility Strategist | W3C AC
> Representative
> Deque Systems - Accessibility for Good
> deque.com
> <https://urldefense.proofpoint.com/v2/url?u=https-3A__gcc01.safelinks.protection.outlook.com_-3Furl-3Dhttp-253A-252F-252Fdeque.com-252F-26data-3D02-257C01-257Cjennie.delisi-2540state.mn.us-257C2a94ca2523bb46a0bdd208d6a321fd94-257Ceb14b04624c445198f26b89c2159828c-257C0-257C0-257C636875763714637908-26sdata-3DGG0O3iMQp-252F8PHf6p8EWzegAcg-252FBpQuuSttIJLwi6EbA-253D-26reserved-3D0&d=DwMFaQ&c=RoP1YumCXCgaWHvlZYR8PZh8Bv7qIrMUB65eapI_JnE&r=b-9TIC95K-nLEKIDibNXAN_FKV-iXhLlAW2Zc3ebV_c&m=_p7Yxrl6Zp2sk72o5dwgmlwhjQ4eLo4fUWPVLEoAbk0&s=Hvm23HBHHOyFc84ojLoyEOH_dH9_6VudfqT7rEX1dGM&e=>
>
>
>
>

-- 
*​John Foliot* | Principal Accessibility Strategist | W3C AC Representative
Deque Systems - Accessibility for Good
deque.com


image001.png
(image/png attachment: image001.png)

image002.png
(image/png attachment: image002.png)

image003.png
(image/png attachment: image003.png)

image004.png
(image/png attachment: image004.png)

Received on Thursday, 7 March 2019 22:09:32 UTC

This archive was generated by hypermail 2.3.1 : Thursday, 7 March 2019 22:09:33 UTC