scribe+
Suji: Introductions, work for Google, interested to learn about ACT
Wilco: Worked on a bug Tom reported in mapping IBM implementation
Daniel: GitHub review from me
<Helen> https://github.com/act-rules/act-rules.github.io/pull/2047
Helen: Work for Teslio, focus on
manual test rules. I created a PR for updating the about
page
... Talking to Chris about what manual rules to start
writing.
Todd: Might be absent for a while for health reasons.
Tom: Working on the iframe rule,
got review from JY. Then I'll pick up ARIA rules next.
... Will be out next week.
Kathy: Worked on label in name rule. Plus I'm working on intro to GitHub guidance for ACT.
Trevor: Working on stateful rules. Will show today
https://www.w3.org/community/act-r
<daniel-montalvo> scribe+
<daniel-montalvo> Wilco: Acction item to reach out to Mrk Rogers
<daniel-montalvo> ... It was about the descrriptiveness of things
<daniel-montalvo> ... They seem to be using machine learning techniques to ensure which techniques work and which do not
<daniel-montalvo> ... This is the first time I saw these applied to ACT
<daniel-montalvo> ... They are also using ACT cases to train their statistical models
<daniel-montalvo> ... This is something we may or may not want to allow
<daniel-montalvo> Trevor: They must have more training data
<daniel-montalvo> Wilco: Yes. I am happy to park this for now and consider it a consistent implementation
<daniel-montalvo> ... We do plan to update ACT Rules Format to be better at these kinds of tests
<daniel-montalvo> Trevor: Would it be possible to ask him to run an experiment to see if he gets the same results without ACT rules?
<daniel-montalvo> Tom: It came up because we did not have enough implementations
<daniel-montalvo> Kathy: I am happy to do Trusted Tester stuff, whatever helps
<daniel-montalvo> ... How do we know wihch data is used?
<daniel-montalvo> Wilco: We don't, there is certain level of trust there
<daniel-montalvo> ... We have a mechanism for the public to report if things look suspicious
<daniel-montalvo> Trevor: Would we allow test cases that are not 100% right?
<daniel-montalvo> Wilco: We may
<daniel-montalvo> ... Are we happy to leave it as-is until we have a conversation about training on ACT data?
<daniel-montalvo> Trevor: It is still pretty rare, so I think we are good
<daniel-montalvo> Wilco: Do we want to move forward with this rule with this implemenation only or we need to wait for a second one?
<daniel-montalvo> Kathy: I can include Trusted Tester results if that helps
<daniel-montalvo> Tom: The implementation shows that there is a process that works, whether it applies or not still needs to be a human decision
<daniel-montalvo> Wilco: Proposal: Move forward and add Trusted Tester results
<daniel-montalvo> Trevor: My guess is training will take just a few minutes
<daniel-montalvo> Kathy: Is that something that the tool owners know how to do or we need to provide instructions for them?
<daniel-montalvo> Wilco: They should know how to do this
<daniel-montalvo> Tom: When you onboard a tool it is good to give some heads-up
<daniel-montalvo> scribe-
<trevor> https://github.com/act-rules/act-rules.github.io/discussions/2046
Trevor: This rule has become a
testing round for different things we want to try. I put this
into a discussion
... How this works, I have the rule at the top, and threads
below it for different pieces
... Please take a look at this over the next week, and add
comments on the different threads.
... One question I do have, the code of these is getting a bit
more complicated, so I'd be interested if this is still
understandable.
... Would this be fine, do you have suggestions?
Helen: Don't abbreviate
"AT"
... It's fine for me because there's a description
Trevor: I would appreciate some feedback on this as well
Wilco: I can put the examples in a codepen and add links
Tom: Wilco commented last time
about triggering scripts from an event.
... Would it help to have some event for this?
Trevor: For the first example I
added an actTfNextAction event, so that a tool can trigger the
event it can run the steps.
... It exists for this example. For the log example I don't
have any event like that. I've thought about a "tool ready"
event, but for this it just puts a message every 3 seconds.
Tom: We might need to put a
property in the page so tools can know the event is there
... I think as long as the first event returns something other
than true it's okay.
Trevor: So try and see if it comes back? We might be able to do better. It's a fair point.
Tom: If there's some property it could let tools skip those
Kathy: In previous discussions we talked about modals?
Trevor: This is on status message. I haven't written the modal rule.
Kathy: If a status message appears in a modal, would it still be applicable?
Trevor: I don't think it would
be. I'd count that as a change in context.
... There is something in here that prevent modal from being
applicable.
Kathy: Part 3 of this discussion.
<kathy> https://github.com/act-rules/act-rules.github.io/issues/1619
Kathy: In a previous meeting we
discussed if expansion of abbreviations are acceptable in the
accessible name
... Conclusion was that expansions of abbreviations would
fail.
... We also discussed if hyphenations to adjust screen reader
correction would be accepted. We decided it would not.
... The next question is on symbolic text. For example the
capital letter X meaning "close".
... In the criterion, symbolic text would not be applicable,
but it was included in the rule.
<kathy> https://github.com/act-rules/act-rules.github.io/issues/1989
Kathy: The rule currently says
with regards to symbolic text in the expectation that it should
be contained, except for non-text content.
... Because there is no expectation they could essentially have
any result.
... Where I'm stuck is that we have no expectation, and I'd
like to write one.
Wilco: I think the exception in the expectation is meant to cover the symbolic text scenario
<Suji> https://www.w3.org/WAI/WCAG21/Understanding/label-in-name.html
https://www.w3.org/WAI/standards-guidelines/act/rules/2ee8b8/proposed/
Kathy: We don't have any expectation for symbolic text characters
Wilco: We put this in the expectation, because non-text content is not objective we're not allowed to put it in the applicability
Kathy: I think this is confusing.
Wilco: How about we add a note to passed example 5 saying the accessible name needs to be tested under SC 1.1.1
This is scribe.perl Revision VERSION of 2020-12-31 Check for newer version at http://dev.w3.org/cvsweb/~checkout~/2002/scribe/ Guessing input format: Irssi_ISO8601_Log_Text_Format (score 1.00) Default Present: thbrunet, Helen, kathy, trevor, Wilco, ToddL, Suji Present: thbrunet, Helen, kathy, trevor, Wilco, ToddL, Suji No ScribeNick specified. Guessing ScribeNick: Wilco Inferring Scribes: Wilco WARNING: No meeting chair found! You should specify the meeting chair like this: <dbooth> Chair: dbooth WARNING: No date found! Assuming today. (Hint: Specify the W3C IRC log URL, and the date will be determined from that.) Or specify the date like this: <dbooth> Date: 12 Sep 2002 People with action items: WARNING: IRC log location not specified! (You can ignore this warning if you do not want the generated minutes to contain a link to the original IRC log.)[End of scribe.perl diagnostic output]