Meeting minutes
New slides added to the WCAG 3 overview deck
W3C Breakouts Day: Update on COGA proposals
There are only 4 possible time slots for all the breakout sessions, not just COGA's.
There are only two slots on March 25 and two slots on March 26, and none are convenient for North America.
We chose a time that works for Lisa and Rashmi
We will go with this time slot on 26 March
10: 00 to 11:00 UTC
12pm Israel time
3: 30pm India time
6am Boston time
<Lisa> questions as we go:
<Lisa> -use needs? do we collect them in a place
<Lisa> -intergrate ai into the process?
<Lisa> -intergrate functional needs
<Lisa> how to give people time to review
Review the schedule for the next few sections
April 2nd conflicts with Passover
Let's see who can attend on April 9 (last day of Passover)
Lisa and John assigned themselves "Process and Task Completion" and scheduled the discussion for April 23rd.
We are looking for someone to volunteer to review an unclaimed section and be ready to discuss it at the April 16th meeting.
New slides added to the WCAG 3 overview deck
Breakout Sessions revisited
We are reopening this discussion.
Lisa is not comfortable with consolidating into one session.
Lisa I am not comfortable with selecting a time that is good for me and bad for others.
Voice systems
Wayfinding
Supported decision-making
Online safety and wellbeing (algorithms and data)
Should we wait until TPAC in the fall?
kirkwood: Could we introduce at the Breakout Sessions and then discuss in depth at TPAC?
Lisa I am equally uncomfortable with either.
kirkwood: I feel like we should have a prerecorded presentation on this.
Lisa: We could record the whole thing the day before at a normal time and then press play on Breakouts Day.
Lisa: And at the end Lisa would field questions.
kirkwood: I would feel comfortable coming in at the end to field questions.
Lisa: Do anyone know how to pre-record and do the captions?
kirkwood: Can we record a zoom session?
Lisa: We need captions.
ACTION: item: Julie to ask Ian if we can pre-record and get help doing the captions?
julierawe: It seems like a lot of effort for a group that is already stretched thin.
kirkwood: But it might be worth the effort to have a "canned" presentation.
Lisa: I think we need to survey the research group.
Lisa: 1 option is to back out.
Lisa: 2nd option is to divide in two — one session at night for Americans and 2nd session during the day for Israel and India.
Lisa: 3rd option is Lisa presents all four in one session.
Lisa: 4th option is trying to pre-record with several COGA members.
We'd need to make clear we don't know if the organizers would allow the 4th option.
ACTION: item: Lisa will survey the research group members.
Lisa initial review/discussion of "Error Handling"
Lisa: I created a tab for general issues for GitHub, such as why aren't assertions requirements.
julierawe: Will we schedule a separate time to discuss the general issues? Or during today's meeting?
julierawe: Len and I can provide background that could help inform issues we submit about the general structure for WCAG 3
Lisa: We may have time today to discuss general issues.
Lisa described the table she created for each provision in Error Handling.
Lisa created two tables, one for each guideline in her section.
Lisa created a section on user eneds
Lisa drafted issues using the COGA template for github issues
Lisa: I merged in Jennifer's comments
Lisa: I don't think COGA should spend time suggesting specific wording. We're doing a holistic review.
Lisa: I really focused on the user needs, less so on the functional needs.
Lisa: I found it useful to focus on the user needs row by row.
Lisa: The table gives me room to write notes and then writing the GitHub issue forces me to clarify.
julierawe: I agree. I did not understand the notes in the first row of your table, but I understand the GitHub issue you drafted.
Lisa is going through the table row by tow.
Correction to the minutes: row by row.
The group reviewed issue #2.
The group also discussed issue #1 to clarify whether we're saying error messages need step-by-step instructions.
LenB: Does an error message need step-by-step instructions? Or just a clear message on what to do to resolve the error?
Lisa: We want the template to say "User need missing or not adequately addressed" and then we also call out "Related user needs from Content Usable"
Lisa reviewed how Eric approached reviewing his section including using AI (NotebookLM) to help think about what was missing.
Lisa: The AI review found several issues.
Lisa: There were additional user needs that were missing.
Lisa: But NotebookLM found things we might not have found.
Lisa described how she took a different approach for "Error Handling"
Lisa: I copied from Julie's presentation where else to look in WCAG that might be related to "Error Handling."
Lisa: For anyone reviewing, it's here. It's all in one place.
Lisa: I also checked the editor's draft to see if there are any exploratory provisions (which are not included in the working draft).
ACTION: item: Julie will consolidate the two Thursday meeting invites into one meeting invite.
Apologies to folks who had trouble joining the meeting late.
Apologies again to folks who had trouble joining today. We will simplify/consolidate the calendar invites for future weeks.
Lisa is going over the "Error handling" tab: https://
Lisa: Any questions about how Eric approached reviewing and how I approached reviewing?
Eric: There were some things where I could not find research supporting what we're suggesting.
Lisa: I did not get that far when I was filling out my templated issues.
Lisa: The review process is a work in progress.
Lisa: We want to keep in mind any issues with the template.
Lisa: I made one change to the template: We have "User needs missing or not adequately addressed" and we added another section for "Related user needs from Content Usable" that are copied directly from our COGA resource.
<Jennifer> + to developing something like a prompt library
julierawe suggested adding a tab with suggested AI prompts
Lisa: Does anyone object to using AI as part of our review process?
kirkwood: I think we should note when it's used.
kirkwood: As long as we know AI is used, then we can review for hallucinations, etc.
Jennifer: I think using AI for first draft is great. I also wonder if there is an opportunity for creating guidance on what we are feeding into the AI tools, so we can be consistent about that.
Eric: One thing about NotebookLM is you can set boundaries and it will only review using those sources.
julierawe: I often make many adjustments to something AI drafted
Eric: My "Animation and movement" tab includes the prompt I used.
<Jennifer> +1 to feeding the template
julierawe: I suggest we add to the prompt the template we want to use for GitHub issues
julierawe suggested we add to the prompt a request to generate a list of user needs from Content Usable
julierawe Then we could look for what is missing from the list.
Lisa: I wouldn't use it as a starting point. I would go to our research modules and Content Usable. And then I'd ask AI.
Lisa: Do we like the idea of making a table for each guideline/subsection and have a row for each provision?
Eric: I think the tables are helpful.
Lisa: Any objections?
Jennifer: I'm a big fan of the tables.
Jennifer: I wanted to get clarification on what to do if there is more than one person reviewing a section. Would we have a column to denote who is saying what?
Lisa: We only want one row for each provision/requirement. We could put in "Lisa:" in the notes section of each row.
Jennifer: If there is a lot of overlapping feedback, then I could add "+1"
julierawe noted some of the biggest section will have very long tabs in our doc because there are so many provisions in those sections.
julierawe but I don't see a way around this.
Lisa noted the importance of creating separate tables for the separate subsections.
<Jan> +1 to the spreadsheet idea
Lisa: Feel free to think about how to improve the process.
Lisa: I also like starting each tab with a background section
Lisa: To start with the subsections, related provisions that live elsewhere, and if there are any additional exploratory provisions.
Jan: I like that but am confused about what would go in the "General issues" tab.
Lisa: One general issue is with the template for assertions. There is a lot of general stuff and that makes it hard to focus on what is specific to that particular assertion.
Lisa: The assertion template could be streamlined.
Lisa: This would be true for every single assertion.
Lisa: Another general issue is how requirements are separated into core versus supplemental.
Lisa: We can let everyone know when the "Animation and Movement" issues are ready for final review.
julierawe: What is the process for that final review?
Lisa: It goes to the group for people to put in their comments and then we have a week for everyone to put in their comments, and then we work through those comments together.
Lisa: The first hour: Let's say everyone has had a week to put in their comments on "Animation and Movement. So the first hour we'd do initial review of "Error Handling." Then I'd have homework to write up the issues. And we'd spend the second hour going over Eric's comments.
Lisa: With Eric, he found initial issues and we found gaps so he has drafted more issues.
Lisa: For mine, we haven't had the full initial discussion of the guideline and we haven't looked for any gaps.
Eric: My suggestion is the first thing you present is a table and that's all you present.
Lisa: Maybe we need the notes to be written more like issues so other people can understand.
Lisa: I was also thinking one issue per provision.
julierawe: Do we postpone Text and Wording on the 19th? So we can talk more about Error Handling? And what about COGA task force review of "Animation and Movement"?
Eric: I could send out the animation issues today.
LenB: Jan and I will meet tomorrow and discuss these two working models. I think we can get a good chunk of "Text and Wording" done by the 19th.
For breakout sessions, we are requesting:
25 March, 1-2pm UTC
9am Boston time
3pm Israel time
6: 30pm India time