[Conformance] Minutes from 25 February

Minutes from the Silver Conformance Options subgroup teleconference of
Thursday 25 February are provided here.

===========================================================
SUMMARY:
*            Initial discussion on two issues posted to github on the Silver FPWD.
*            Initial discussion of several new use cases, including one for
              component libraries and another with XR implications.
===========================================================

Hypertext minutes available at:
https://www.w3.org/2021/02/25-silver-conf-minutes.html

===========================================================

   W3C

                                                                                                            - DRAFT -
                                                                                               Silver Conformance Options Subgroup

25 Feb 2021

   IRC log.

Attendees

   Present
          Bruce, Janina, jeanne, Jemma, JF, John_Northup, JohnNorthup, PeterKorn, sajkaj, sarahhorton, Wilco, Wilco_

   Regrets
          Jemma

   Chair
          sajkaj

   Scribe
          JF

Contents

    1. Agenda Review & Administrative Items
    2. Assigned github issues
    3. Use Cases Discussion (Continued)

Meeting minutes

  Agenda Review & Administrative Items

   JS: new item - silver has begun to assign comments. many have come in

   some have come to this sub-group for our feedback. 2 in particular

   we also have several new use-cases to review

   JN: I added some new use-cases - or at least partials

   JS: we also have another use-case from Wilco. He hopes to join our call shortly

   zaakim, next item

   <sajkaj> https://github.com/w3c/silver/labels/Subgroup%3A%20Conformance%20Options

  Assigned github issues

   link https://github.com/w3c/silver/labels/Subgroup%3A%20Conformance%20Options

   JS: wonder if we should read through these?

   JS: yes, especially with the first one(s)

   JS: reads out first comment "On Scoring"

   TRiaged this for this sub-group based on final paragraph: However, we hope the WG will clarify whether the different thresholds or other classifications are meant to indicate that some outcomes are more important
   than others. Such indications, if there are any, might be valuable for authorities when evaluating exceptions due to a disproportionate burden.

   JF: believe not all reqs are 'equal' yet they are scored on a flat scale

   JS: may not be a violation, but impact on "minimum difficulty"

   see this often in Adnroid apps

   if you take the time and examine the buttons, you can label them, but they aren't labeled natively

   Jeanne: red this as a burden on the content creator/site owner and not the user

   SH: I read it that way too

   this is adding another layer - 'complexity' (?)... resources to fix

   JS: to the example: can we specify the example of alt="button",alt="button",alt="button"

   SH: figure that would score low on complexity - low burden on author

   <jeanne> +1 to Sarah

   JS: seems to indicate controls, as opposed to an informative image

   JF: notes the different types of images

   JF: asks if actionable images are more critical than informative images

   JF: what question are we attempting to answer? What JF and JS heard, or what Jeanne and SHG heard?

   JS: Maybe we go back and ask for mor specificity/clarity. We see perspectives, maybe ask for examples?

   JF: +1 to returning for more details\

   SH: we could adopt the idea that conformance approach will support orgs that need to make determinations and add use-case

   we know ultimately that orgs will need to be making those kinds of decisions - examples would be beneficial

   Jeanne: it would be helpful to have some ideas written up

   if we had examples, we could show them to other stakeholders for feedback

   Jeanne: found this interesting as something we never considered

   what are the pros and cons of this? we don't know
   ... so we should ask for examples

   JS: Will respond on our behalf and request more examples/illustrations

   JF: URLs or usecases

   Jeanne: and why that is important to them

   Jeanne: this comes out of the presentation to Access Board/Trusted Tester

   [Jeanne reads second issue]

   <PeterKorn> I'm so sorry I'm coming in so late. Wasn't a great night...

   JF notes that bug trackers like JIRA have 5 levels of severity: Blocker, Critical, Major, Moderate, Minimal

   also notes that tools like axe-core uses 5 levels of severity as well

   JS: the only thing that troubles me is that it adds complexity to the scoring mechanism
   ... we're no longer an itemized list, but instead a table
   ... like the idea of making it part of the normal bug-tracking process

   PK: hars the concern about adding the complexity, but like how this dove-tails into principles
   ... perhaps we take this on as a 'test' - try applying it to our test examples
   ... but move to levels of severity

   But there is also the idea of the "spoons" concept... adding up 'criticals' ... exceeding your "spoons" limit

   +1 to enumeration of "how many issues"

   contrasts 40 major bugs versus 100 minor bugs [sic]
   ... may be a good way of capturing the cumulative friction issues

   PK: we had the prior example of a 200 page document that lacks headers: lacking one page versus lacking all pages

   <Zakim> jeanne, you wanted to say that there is a serious structural problem with criticality as in the example because it penalizes cognitive disaiblities more than sensory disabilities

   Jeanne: the example provided worries me because it seems to perpetuate the structural bias problem

   where people with sensory disabilities are generally considered higher priority because there is a workaround

   Jeanne: get lots of questions on this - how?

   and we're not sure

   <PeterKorn> When do we need to respond to these GitHub Issues?

   SH: thinks this is an interesting perspective to explore more deeply
   ... think one of the things that is interesting so far is notion of critical errors apply to outcome, and also apply the across functional categories
   ... its a coarse measure, and i'm in favor of breaking things down, along with functional categories and user needs

   JS: thinking in terms of follow-up, respond to issue that this is interesting and we're still working on it

   PK: until we have better edit control, want to avoid pointing folks to Google docs (fear of over-writing content)

   JS: OK, will respond, but want to move to next item before end of call

  Use Cases Discussion (Continued)

   JS: would like to ask Wilco to expand on his contribution

   <PeterKorn> https://docs.google.com/document/d/1GyUYTnZp0HIMdsKqCiISCSCvL0su692dnW34P81kbbw/edit#

   WF: added 2 - one I mentioned last week plus a new one

   WF: first one is component library, and want to express the a11y of that

   and division of roles between author of library and content creator using the library

   JF: notes that design systems are 'higher order' than juust component libraries

   PK: this is interesting - how would this pan out in the real world

   any author can create inaccessible content, even when using accessible components

   PK: does there need to be an explicit GUI library/component, or might this be like WCAG 2 Optional conformance claim

   :PK how does that feel/sund?

   WF: not sure
   ... feels meaningless then to say "I have an accessible component library"

   can we say something about accessible pieces used to build larger content?
   ... is there a way to avoid that?

   PK: see where you are going
   ... notes the difference between unit testing and integration testing

   PK: talks about all of the issues related to an accessible button - there is a class of requirements, but the 'text' in the button may be a massive fail
   ... we might have a large class of components. Is there a level where only components need to pass

   JF: believes that use-cases like CMSes and 'templates' taht will want to be able to make claims

   PK: thinking about this - wikis

   <jeanne> I believe that WCAG3 FPWD has included the "not applicable" and should be able to handle components.

   PK: the way to achieve is to pair it to a level that doesn't require human asasessment

   JS: wonder if we can pause here?

   JN: added some ideas involving 3rd party content - one was a LMS associated to a college.

   with multiple authors adding content

   wonder if there is some kind of algorithm - similar to flesh-kincaid or similar - but feedback in real time

   JN: other examples is "virtual tours" on real estate site

   may add some sparse text that describes apartment

   <Zakim> jeanne, you wanted to talk about 3D

   Jeanne: XR sub-group have been discussing a similar concern: text equiv in VR/AR (XR)

   WR Subgroup are talking about tying it to components in XR


    Minutes manually created (not a transcript), formatted by scribe.perl version 127 (Wed Dec 30 17:39:58 2020 UTC).



----------------------------------

Janina Sajka
Accessibility Standards Consultant
sajkaj@amazon.com<mailto:sajkaj@amazon.com>