Meeting minutes
Sub-group check-in
Shawn: Updates on Friday. Updates from others?
<Makoto> Results of the Attendees Survey - Japanese Webinar on FPWD https://
Makoto: Want to share a summary of my webinar.
… URL for an English translation of Japanese comments.
<Makoto> This is the version Maya was working on translation. https://
Makoto: I did a webinar on the FPWD on Feb 19th. Had some 150 attendies. I got 36 responses on the follow-up survey.
… I asked 5 questions. Do you believe WCAG 3 will be easier to understand.
… 52% answered it would be easier. 38% answered not sure.
Second question; what do you think of the user interface of the documents. 30% answered much easier, 50% answered easier. 19% not sure.
Third question; what do you think of using points? 33% answered great idea. 54% answered good idea.
Question four; How difficult do you feel it would be if Bronze is AA. 52% answered difficult. 28% answered just right.
Last question; Conformance claim if there are some problems. 44% answered good. 27% answered not sure.
… I was interested in how people answered the question. Especially question 3 and 4. These were big changes from WCAG 2, there are pros and cons. I am neutral at the moment.
… But I am concerned about how to create a scoring system. It must not depend on who the tester is. Results should be repeatable.
… According to the survey, people like the approach. There are benefits over WCAG 2 style. On the other hand they worry about how to do it in reality.
… Might say no after they have concrete examples of guidelines and outcomes.
… About 70% answered (very) difficult with bronze as equivalent to level AA.
… WCAG is an international guideline. There are many regions where there is no obligation to make content accessible.
… Many people left a comment saying that if bronze equals AA, it is a very high level. Not easy to achieve.
… If Japan had a legal obligation it may seem reasonable. In my opinion it is more reasonable to have level A as bronze, and have silver / gold as bonus levels.
<Jemma> written comments from the survey are really great and helpful
Shawn: Very interesting comments, specifically on bronze level and AA.
chuck: Didn't quite follow Makoto's proposal. Worth exploring.
<Chuck> Wilco: I wanted to ask Makoto, is A an appropriate level, or is that too high a level? difference between A and AA isn't that great.
Makoto: In Japan we need to encourage more people to do accessibility.
… If WCAG 3 sets the bronze level equal to AA, it will be too high. If we could set the lowest level lower, more people will start doing accessibility.
… I don't want Japanese professionals to give up bronze level because it is too high.
<Chuck> Wilco: Didn't quite answer question. About single-A level, is that a good target, or is that level too high as well? Would it be better to have a slightly lower level?
<Chuck> Wilco: Is single-A a good baseline for Japan, or should we go for a bit lower for that?
Makoto: Level A is reasonable as the starting point for many Japanese websites.
… We would like to keep that as the lowest level.
Jemma: Want to highlight 3 of the comments. First, we need to use clear language for the spec. Second on a kanji character for color contrast.
Makoto: We want WCAG 3 in plain language. Makes translations easier than WCAG 2.
<Lauriat> Added the described proposal as option 13 in our doc: https://
Jemma: There was a comment what the difference with how-to and method. In Korean this has the same meaning.
<Jemma> my another comment was concern for the color contrast of kanji fonts
JF: Agree with the concerns on repeatability. The other comment, we were going to start with a blank sheet, but seem to still be stuck with WCAG 2's A, AA, AAA levels.
<Lauriat> +1 to JF, I think we can use this feedback as a way to think about Bronze, Silver, and Gold, rather than as a direct mapping of A, AA, AAA.
Makota: People like the way of thinking in the new approach, but care about how we can do it in reality.
Chuck: Regarding how we avoid A / AA comparison. If we don't guide that, others will make the comparison. We need to guide it in the direction we want it to go.
<sajkaj> +1 to Chuck
<johnkirkwood> +1 to Chuck
JF: The scoring mechanism already has a range of achievable score. I think we should work with percentages. A team can go from 70% to 78%, that's measurable progress.
<Jemma> to chuck's comment
<Zakim> Lauriat, you wanted to note one aspect of A vs. Bronze, etc.
Shawn: Even if we end up with essentially the same model, one benefit we get is that people will regard bronze as knowing it can be better, as opposed to "A" which is many cultures is the highest level.
Jemma: We can not totally disregard the A / AA / AAA system. It is a good reference point. We need to change the conceptual model.
Sajkaj: Agree with Chuck, we need to define the comparison. We can go beyond that. One way we might look at defining conformance might be on a achievability scale.
… If you don't have the framework in place to fix automated, you shouldn't worry about bronze. Wouldn't make that normative, but somewhere we should acknowledge that.
Shawn: Added an option in the BSG doc, with a link to the survey.
<Jemma> +1 to "achievability scale" by sajkaj
chuck: The world is going to define the relationship between A/AA/AAA and Bronze/silver/gold. We may want to consider guiding that direction.
<Jemma> great point, Chuck!
Shawn: Agreed. Use that to talk about the differences.
<Makoto> Thank you all for your questuons and comments. I really appreciate it. Alt text subgroup will have Maya as a new member and will start working on answering comments we got from the first public working draft.
<Lauriat> https://
Shawn: As of minutes ago, we've solidified as Thursday April 29th as an all-day working session
Options for Bronze, Silver, Gold (resuming with option 8)
Shawn: Resuming bronze/silver/gold discussion from option 8
<Chuck> Description Bronze (Low subjectivity) Guidelines that are not subjective Subjective guidelines on a rudimentary (“everyone can agree”) level
<Chuck> Silver (Slight improvement in non-subjective. Bulk of subjective work falls here.)
<Chuck> Lower margins of error for large sites on Guidelines that are not subjective
<Chuck> Subjective guidelines require evidence: a written explanation including some evidence such as: User testing/Remediation/Re-testing report, User testimonies, ???
<Chuck> Gold (Significant improvement in non-subjective.)
<Chuck> No margins of errors on Guidelines that are not subjective
<Chuck> Non-subjective guidelines require efficacy evidence: a written explanation including some evidence such as: Successful user testing, User testimonies, A published schedule of AT-testing by people with AT-relevant disabilities and publicly-available issue triage
<Lauriat> https://
<Chuck> Pros [none noted in the meeting of 23 March 2021]
Shawn: Any questions about the description?
<Chuck> Cons Published documentation: Where, who, accessibility of those statements? It would be a very large document for a large site.
<Chuck> Why would we tie subjectivity of a guideline to the accessibility of a site?
<Chuck> It is turning Bronze, SIlver, Gold into a redux of A, AA, AAA.
<Chuck> Machine testing only goes so far. At some point there is an evaluation if it is done correctly, which will always have an element of subjectivity.
<Chuck> Would constantly need updating due to advances in automated testing which would require changing the rating of a guideline. That would not be acceptable to regulatory bodies.
<Chuck> (same as Option 6) It does not provide a structure that does more to provide fair treatment of all disability needs. It is too easy in this structure to assign the ...
<Chuck> hard-to-measure needs of people with low vision and cognitive disabilities to Silver and Gold levels, where they will be considered optional and not be met. This is a civil rights...
<Chuck> issue. Unless we can put in structure to ensure that the hard-to-measure needs of people with disabilities will be addressed at bronze level, then we should not go forward with this proposal.
<Sheri_B-H> one more con: public bug databases are super problematic because of confidential information
<Chuck> Wilco: There's a bunch of comments on automation. Is there... are people thinking that subjectivity and automation are somewhat equivalent?
Shawn: There was a proposal saying that an initial level could be automated.
Wilco: Think that was proposal 11
Sajkaj: Yes, that is a proposal, necessary but insufficient level to start with.
Jemma: Surprised there are no pros for this option. It is easy to understand
… Second item; for example second item. We're talking about the subjectivity of the testing.
… I don't understand the last con.
Shawn: Seems like it is taking BSG and mapping it to the types of tests, as opposed to mapping it to how accessible the site is.
… Testing does not necessarily map to an improved user experience.
Jemma: This option suggests that at a higher level, you have to do more testing.
Shawn: Seems to map it to the types of tests. Today there are some tests much easier to do than it was 10 years ago.
Sheri: I think you said oen of the levels required a public bug database?
Shawn: Yeah, public documentation.
<JF> +1 to Sheri
Sheri: That is super problematic. Gasoline on a fire. Confidential information is discussed in private. Could never be public.
Shawn: I agree, also because it's getting more into the area of a maturity model, but for a particularly shaped organisation
<sajkaj> +1 to Sheri
<Lauriat> +1 to Sheri
Sheri: We have an open source design, but have a closed JIRA system with much more detail.
Shawn: It's more of a way to communicate with users. That seems odd to tie to a conformance level of a product
<Chuck> +1 to Sheri (with chair hat off, as Oracle)
JF: I work with many companies where we sign NDAs. Am concerned about anything that would require public admission of problems.
Shawn: I think we worked through that; saying we would not say it needs to be in a format / place.
… Any -1's to saying we've discussed this?
… Seeing none.
Shawn: Second; there is a difference between subjective and non-ambiguous.
<Chuck> Wilco: That sounds like my point :-)
<Chuck> Wilco: I wonder, how do you even determine subjectivity of a requirement? Something like "title describe page" and language of page correct? Do both have some level of subjectivity?
<JF> +1
<Chuck> Wilco: Which one is more or less? That's a tough call.
Shawn: Think the proposal is addressing unambiguity, not subjectivity.
Jemma: Subjectivity here is for example about whether something has a title.
… As you see from the option, bold has more testing requirements than silver and bronze
Shawn: Would that mean that taking title of a page. So bronze would be if you have a title, and how relevant it would be would be at silver level?
Jemma: Yeah, I think so.
… Haven't talked about different options for scoring systems. This is about how we're going to deal with subjectivity.
… Trying to apply these options to a new scoring system.
Shawn: One thing to get out of this is to avoid ambiguity, and structure the test so it supports consistency in making subjective calls.
Sajkaj: Checking if an image has an alt seems to me an easy thing you have to achieve. You would want to make sure you have that. Breaking it up that way might be useful.
… It doesn't take you where you need to go, but starts you in the right direction.
… Even when making subjective judgement, we should get more systematic knowledge, so we can better avoid ambiguity. Especially if this effects multiple categories.
Chuck: One of the challenges Oracle came with, for structured content, how many heading should there be. We came up with different numbers from different people.
… 8 to 10 is a huge range. I think that helps describe the issue.
<JF> +1 to Chuck
<Jemma> +1 to chuck
<johnkirkwood> +1 to Chuck
<jeanne> +1 wilco
<Chuck> wilco: Can we do that in the scoring system, rather than in the conformance level? Maybe we lost it along the way. A scoring of 1 or 2 through automation. 3 or 4 may or may not include subjective that are harder to automate.
<Chuck> Wilco: That may be a way to get the benefit w/o tying to a hard conformance level.
<Jemma> +1 to Sheri about automated testing
Sheri: We're going to need to provide some blessing of which automated tests constitute core tests that can be completed. Not an area we should go, and might stifle evolution of automated testing.
… Can't see how to make that work
Shawn: In favour of automatable in theory, instead of automated in practice.
<Zakim> jeanne, you wanted to point out the civil rights issue precludes using option 8
Jeanne: I think the civil rights issue precludes of from using option 8. This is saying that disabilities difficult to test for are optional again.
+1
<johnkirkwood> +1
<JF> They ARE addressed at "Bronze" and WCAG AA today
<jeanne> JF, WCAG2 is a problem in this regard. It must be addressed in WCAG3
<Jemma> jeanne. can you address this issue to Option 8?
<jeanne> If only the easy to test disability needs give bronze level, then people with cognitive disa iblities and low vision will not have their needs addressed at Bronze level, only silver or gold. That is not treating all disabilities equally. It is a structural problem with WCAG2 that we MUST fix.
<Jemma> ah... i got you.