Minutes 2021 01 14 - act-rules/act-rules.github.io GitHub Wiki

Present: Wilco, Carlos, Jean-Yves, Shadi, Jon, Sameera, Emma Scribe: Jon Avila

Final call

https://github.com/act-rules/act-rules.github.io/issues/461

<Wilco_> https://github.com/act-rules/act-rules.github.io/pulls?utf8=%E2%9C%93&q=is%3Aopen+is%3Apr+label%3A%22Final+call+2+weeks%22+

Wilco: 1 rule in final call in two weeks list. iFrame with negative tabindex

Wilco: put this into 1 week final until it's resolved. Updated expectation.

Rules ready for W3C publication

https://github.com/act-rules/act-rules.github.io/issues/1120

Wilco: list of all the rules that are making their way onto W3C website. A couple of change requests.

Wilco: would like to get these resolved?

Carlos: 29th due date

Wilco: Image filename is slightly complicated? Taskforce is not sure about this. Wondering if it should be dropped.

Jean-Yves: was written to a few cases like content managers that just add file name.

Jean-Yves: directly human review. Would be only value compared to full image accessible name. Don't have strong feeling to keep or not.

Wilco: Taskforce is happy to bring to AG if we want it. Does anyone mind if we deprecate as it's superceeded b main image rule.

Jean-Yves: rule about image file name being same as accessible name of the image.

Emma: seems reasonable to deprecate the other one.

Wilco: Jean-Yves create pull request to deprecate and that would go through final call. Mark deprecated in title.

Wilco: on to auto complete.

Aron: headers attribute in ready for W3C. Should I send now?

Wilco: Yes, start moving stuff over to them and let them decide how to handle work load.

Wilco: waiting for task force to resolve. Have been surveyed and need to open some tickets.

Wilco: ARIA has valid attribute has open issues - need to look at.

Wilco: is anyone interested in taking anything new on? Given that we have substantial list?

Sameera: wouldn't mind helping - something simpler.

Wilco: will put your name on menu item on and will reach out to you by email.

Wilco: added 4 new rules to W3C list.

Should every foreign word be in a lang attribute

https://github.com/act-rules/act-rules.github.io/pull/1521

Wilco: is this blocking a rule?

Jean-Yves: may not be a blocker to a rule.

Wilco: Were righting a rule on 3.1.2 language of parts and got into conversation of does every word need to have lang attribute. Does it need to be about paragraphs or larger things?

Wilco: Does a single word in Spanish need to have it?

Jean-Yves: changed to each word after initial approach of primary page language. It is an open issue with WCAG. Tried to sum up arguments.

Jean-Yves: Summary - says passage or phrase and doesn't say word - can be very disturbing for screen readers. Argument for is that the correct language - full paragraph in understanding when single words explain when and when they do not need. In many cases single words need to be produced different for undestandability. For example, language menu. Lastly interpretation - single words might fall under exception if they don't need to be [CUT]

Jean-Yves: what do we want to check in the rule?

Emma: Could be subjective to when you should or should not mark it as in a foreign language.

Emma: would be really helpful when it's a word that is a common word in other languages. English pulls words from all of the place that are pronounced correctly.

Aron: Example of Welsh language selection - a phrase could be a simple word. Then a single word would need to have the specific language.

Aron: Phrase could be a single word

Emma: I agree with that.

Carlos: Doesn't mean every single word needs to be tagged.

Carlos: In favor tagging each word in the correct language for pronunciation.

Emma: multiple words together would be a phrase and the only the whole phrase would need to be marked.

Wilco: Not every single word but it could be individual words.

Jean-Yves: rule is other way around. Looking at lang first.

Wilco: what does that mean for the rule change.

Emma: Need exceptions for words common in language like cliche.

Emma: Example paragraph with Dutch word should have span with language markup

Wilco: Do we need to define - or could we say check each block of text and mark sure language is correct for each block. That would leave more testing to be done but that would capture broad strokes?

Jean-Yves: easier to target language attributes than paragraphs

Jean-Yves: Is the one you put correct - different problem if you have not used it when you should have.

Emma: Does the language describe the bulk. A very different test would be each language marked up.

Wilco: sounds like we want 2 rules but we are writing one now.

Wilco: Check that body language defined and make sure it corresponds with the most used language in the phrase of the language attribute.

Carlos: A paragraph could have both Spanish and English and you could pass the first rule but not the second rule.

Aron: only language needs to be specified - doesn't say it needs to be appropriate.

Wilco: don't want to get into that rabbit hole. That is the intended meaning.

Aron: would agree that it's a failure based on reasonable assumption that the one English word would need markup unless it's common in that language.

Wilco: Spanish paragraph - English sentence - do you want this rule to address that or have a separate rule?

Jean-Yves: Target every sentence in every paragraph but might still be difficult.

Carlos: sentences can mix words from different languages.

Emma: think need 2 rules. Correctly predominately matches the thing it's wrapped around. Need another rules that a person goes through page to look at each phrase that is different is marked up.

Aron: We have rule for checking language that could be valid.

Wilco: does anyone object to us moving to applicability for language attribute and most of the language matches?

RESOLUTION: continue rule development with most common language

Organising background citations

https://github.com/act-rules/act-rules.github.io/issues/1531

Wilco: background sections are getting long -Jean-Yves has put info on how we can organize the background better.

Wilco: some alternatives where sub lists can be used with related rules and linked resources, etc. like HTML

Wilco: That is option 1. Option 2 is to preface each link with what it is.

Wilco: Either in the link text or outside the link text. Wilco: Another option with headings.

Carlos: headings option is my favorite.

Sameera: like to go with headings.

Emma: like the idea of headings - was wondering if background section might not be long enough for headings.

Wilco: many will have list with one or two and then headings. Personally I think they are a little short for headings.

Wilco: quite like sublists.

Jean-Yves: sublists is preference.

Aron: I'd prefer headings

Emma: could I suggest 2 headings - combined approach. 1 for background reading and 1 for related rules.

Emma: Could precede link with some context or make it clear what it is from link text

Emma: Background reading could be single list and related rules would be separate list with heading.

Wilco: want to put this past the task force before it's adopted. Can someone write this up in the issues.

What to do of "role attribute has valid value"?

https://github.com/act-rules/act-rules.github.io/issues/1496

Aron: We had discussed in past and did not reach conclusion.

Wilco: wrote rule to map to 4.1.2 name, role, value. Then task force that great - but it applies to user interface components - take 4.1.2 off of it that does not map to WCAG. It needs to map to WCAG.

Wilco: doesn't say you can only use these rules.

what about 1.3.1?

Jean-Yves: if you have role of btn it has a role of widget - usually target stuff that has a role of widget.

Jean-Yves: If role attribute is wrong - we want to catch that

Wilco: assumption to map to 1.3.1 is that everyone element with a role is presented as something. If it looks like button then it need rule.

Wilco: this rule is looking at unknown or mispelled rules -so we don't know if the rule is for a widget or something else.

Carlos: should it apply to only visible or accessibility tree?

Wilco: let's try to map to 1.3.1. We can have those discussions in a review.

Shadi: would be glad to have language resolved.

⚠️ **GitHub.com Fallback** ⚠️