Meeting minutes
Navigating to heading in main landmark; test feedback issue 812
Example we are testing: https://
James: We are navigating forward from a link, there are several commands.
For this one we want to know if the screenreader reads the title of the container. Joe reported when we convey the role, that shouldn't be conveyed and is confusing
James: This is also about modifying the example in APG
Matt: Was this an optional assertion? I assume when you land on a heading that announcing the landmark container is optional
James: I think its a nuanced area, this is also a common web practice, when users press 1 on a webpage and land on a heading, they want to know if they are in the main content
Michael: I think the substance of the question is less about the optional assertion, more about the wording of the instructions being confusing about what we are testing
Matt: I kinda wonder if this true for all the landmark related test, do we want to use the APG as the test case
Michael: I don't have any strong options on that, I think having the word main" and heading" is potentially confusing as we expecting a role to be conveyed. However for screen reader users this wouldn't be unexpected
Michael: I don't have any strong options on that, I think having the word main" and heading" is potentially confusing as we expecting a role to be conveyed. However for screen reader users this wouldn't be unexpected
Matt: I guess this wouldn't be an issue for automation, so I agree with what Michael is saying
Michael: Now that being said I think it would be best to avoid it and have more real world examples
Matt: We do need to rework the Landmark section in APG as it doesn't fit the new design, and was build in day when APG was giant long document, and much the content was duplicated throughout. Now we have new architecture that doesn't require duplication
Matt: I'm not sure if I have a issue open for the redesign of the landmark pages, if I do I need to link this issue to that one. I will add that to my To do list
Matt: Are this is a won't fix and we will address in the future with an APG change? It doesn't seem like something we should put our energy into
James: I'm okay with that
James: I will add a comment to the issue describing the future fix with the apg change and the note that this is expected behavior for screen reader users
Matt: Alright I have this in my to do list
AXE-con plan
Matt: I noticed that the Axe-con deadline was extended to December 13th, which is great. I think we are at a point in the project that we start to share the milestones we've achieved in 2022
Matt: Getting support tables into the APG is something we need to share. Not just that they are there, but when people can expect more to be there. We need to do everything we can to do socialize this
Matt: I will put a abstract together for this. How many people can be a part of a single presentation?
Michael: I can find out
Matt: I feel like there are several dimensions of our work, we need representation from Automation so someone from Bocoup talking about that
Matt: On the APG side, I can speak, or maybe Issac can speak to some of the research and design
Matt: For sure James, and myself. Maybe then some from Bocoup and Michael
James: It may not be able to be myself, it may be Sina
Matt: I'm assuming the slide deck will be due in early February
Matt: It would be good to have some insight into when the recordings will be done.
Michael: It will be live
Matt: Okay that is easier
Matt: I will take the lead on the initial proposal.
Matt: I will try to get it out in the next couple of days, so I can get input from a wide variety of people. I will share with likely participants
Michael: The alternative, could be submitting multiple talks
Matt: I suppose yes. 40 minuets if we are not going deep on anything, just an intro session, seems like a best approach. Should we go deeper?
James: One of my thoughts about to make this presentation positive would be to clue people in the industry into the fact that this project is going on
Matt: I think an intro talk, with a slide or two with a few sentences, that are aimed at people with a higher understanding would be helpful, but overall we want to keep it introductorty. My overall goal is that more organizations see the value in the APG and looking at the tables for their signal when a particular pattern is mature and ready for broad usage. If it becomes this kind of resource industry wide, than the value of the ARIA AT data[CUT]
Matt: increases
Michael: I think demoing some of the ARIA AT demo would be great
Matt: I would love to demo the NVDA Add on, but as a secondary goal. What is our goal there with the automation? Do we want more people who would be exploiting the automation work, to join that community group? Encourage them to join calls.
Matt: I will send out a proposal, then we can talk more as a group about our objectives with specific topics
Getting first set of support tables into APG
Matt: I want to discuss the sequence of events that need to take place, what do we need different people to do and when? Then we can talk about the technical mechanics
Matt: The first thing is decision is about whether or not we have any data right now, that we could propose go into the APG without any more work
Seth: This is question about the content of the candidate data? Not a technical question
Matt: Yes
Matt: If we look at the current candidate data, is on the reports page or the candidate page
Matt: James do you have any suggestions based on where things are now?
James: Lets look at the reports page: https://
James: The command button example looks good
James: The Link Example 1 (span element with text content) is 100% for JAWS, Chrome, and voiceover, 96% for NVDA
Matt: For toggle button did we not come to resolution with apple?
James: No we did not
Matt: For command button, I think we should discuss why NVDA is failing
James: The test was run with a old version of NVDA
James: The test pass for me in chrome
Matt: Okay, but this would require some test work
James: I don't feel comfortable publishing reports that are failing for a bug they have already fixed
Howard: We made a fix in the next release to alphabetize the tests on the report page
Matt: So link can be out test one
Matt: Then we should prioritize how we go after that
Michael: Do we have a scenario for a multiline link?
James: No we don't but we could
Matt: We will start with Link Example 1 (span element with text content) and test out showing data on APG
Matt: For the ones that we have to do some restesting, what is the way forward?
Matt: For the ones we are rerunning with a new version of the screen reader
Seth: The short answer is the test management page offers the opportunity to make changes like this more straight forward. There are ways to achieve this today, however we want Issac to work on designing new ways to perform these tasks
Seth: The steps will be to Add a new version of screen reader, add new entry for the test, submit data for the test queue, and it will appear to supersede the current data
Matt: If we do this, does that mean we need to restest all data with all screen readers to get data on the reports page?
Seth: The reports page shows the latest data we have for everything
Howard: It would show as two different rows in the table
James: What we are envisaging is to add a test plan to the Queue for just JAWS, carry out the testing, then two reports will show for the same test, which is not what we want.
James: This all raises good questions about test versioning
Matt: In the next couple of days, can we test the shortest test on sandbox test updating the test results and see what happens when we rerun the test with a newer version of NVDA, then publish it
Seth: I think this is a task for us
Seth: I think what we are trying to achieve is a little unclear
Seth: Not all of the data we want to show exists on the report page. We can run a test and if we run into ambiquity we can discuss it. Bocoup will take the command button example and run the test with a newer versions of NVDA and get the data to the reports page, then see what it looks like
Matt: we need to know ASAP if there is some manual clean up we need to do when retesting
Matt: The really important thing is that the data in the APG page needs to match the data on our site
Seth: Yes we agree what is on the APG page will be the latest data of everything by default
Seth: We will send an email update by EOD Monday