07 December 2021



Meeting minutes



AG feedback on testable outcomes

Jeanne: We can address Andrew's comment, to be developed on creating a representative sample.

<Francis_Storr> https://docs.google.com/document/d/1onytB133NRc1EXEOVPizcOqeTi70r2JUcba-LgUyKEY/edit#

<jeanne> https://docs.google.com/document/d/1sugAtqie_x1XqHDZo1Im7ftDNllWeRV_ty4PULeoTV0/edit#heading=h.q6kvhdps0qv4

Wilco: Don't think sampling should go into this document

Jeanne: Alright, just want to provide an answer.

Wilco: Andrew's not in AG often, so maybe a response in the survey is a good idea.

Jeanne: I'll edit my comment to include a response.


Jeanne: What to respond to Mary Jo?
… I always thought these should be clarifications rather than definitions.

Wilco: Not sure that works. It's still a definition, even if we call it something else.

Jeanne: You can say "in this application, we mean this"
… Everyone has its own definition of heading. We can constrain what we mean by heading.

Wilco: I prefer to use qualifiers with that "semantic heading" rather than "heading".

Jeanne: The definitions are always really useful when they are narrow, and have examples

Francis: Before you do anything, you need to define what you're talking about.
… having definitions, understanding what they are is super useful and important.
… Otherwise we'll end up in WCAG 2 territory, arguing what a focus is, or what a border is.
… They can be tedious to write, but getting it tied down is really useful.

Jeanne: There is a WAI subgroup working on a glossary.

Wilco: I don't think we need to worry about it too much, as we can change the definition term without changing the definition.

Jeanne: So research existing definitions

Wilco: We already have a sentence saying not to use definitions from existing standards. Should we add a list of standards?

Jeanne: Yes, I'll work on that.

Programmatic language

Francis: I reached out Jason for answers. He didn't know the answers to our questions, nor did he know anyone who could.


Francis: Some questions, when is a word widely enough adopted in a language to no longer need a lang attribute.

Daniel: Most people tend to prefer not to change language unless you're bilingual. Once a word is established, people feel okay.

Francis: In terms of scoring, is this down to the auditor?

Daniel: It's not just whether it is acceptable or not. It's not a very big issue. People can live with these kinds of things.
… Where do we set the boundary on what is and isn't acceptable. What's good in one community might not be for others.
… This might be an issue, if there are words that are not recognisable.

Wilco: Just to put it out there, is it actually an issue if individual words do not have a lang attribute?

Wilco: Do we have an example of a sentence with one word of another language, that we know for sure needs that lang attribute, or it will be a problem.
… Words that have different meanings in different languages?

Francis: yes

Wilco: Better assistive technologies can make this outcome obsolete

Daniel: Most of the time people prefer the weird pronunciations, rather than the dictionary pronunciations.

Wilco: So should WCAG 3 have a programmatic language outcome?

Daniel: There is a case, for example the browser local is different from the language of the screen reader.
… That's an issue, if the page doesn't have a language, there is a guess. It goes to the default language, you may be forced to change the screen reader.

Jeanne: What about if you have a large chunk of text, a paragraph in a different language. Wouldn't a language change be appropriate then?

Daniel: Yes, I speak a little French, for that it would be.
… When it comes to longer texts, there it's more of an issue.

Jeanne: It would certainly be easier if we moved away from individual words.
… One way we could address this is to have a method for screen readers, and say they have to do language detection.
… Makoto felt it important to include things to do for screen readers.

Wilco: It doesn't have to be screen reader, it can be the browser, or an extension in the browser.

Minutes manually created (not a transcript), formatted by scribe.perl version 185 (Thu Dec 2 18:51:55 2021 UTC).


No scribenick or scribe found. Guessed: Wilco

Maybe present: Daniel, Francis, Jeanne, Wilco