Wilco: We don't really know what the relation between Silver and ACT RF is, that's why Jeanne was invited.
Jeanne: Silver will have guidelines in plain language that will be technology neutral. Methods will be associated with the guidelines and will include tests. What we want to do is move away from page scoring to "project" or "product" scoring. Whatever method people use will gain them points, with a minimum requirement on points needed for different categories. All points then add up to a score which will correspond to "bronze", "silver", or "gold" level.
... One thing being discussed is what should be considered "bronze" level? Some have debated that it should be all automated tests and the minimum requirement. We're therefore very interested in using the ACT work and either linking to tests or pulling them in, that I don't know yet.
Mary Jo: For enterprise level apps and large sites, it's very difficult to fully meet a single criteria.
Mary Jo: Automated tests are also never going to be 100% correct.
Jeanne: We're also looking at something like "substantially conforms". If you can show that you've done the due diligence, some think that you should get the points. For example, if 90% of your images have text alternatives, you get some points. If only 80% have text alternatives, you get less points.
Wilco: A lot of things in the WCAG make it hard to enumerate content, for example counting the number of images on a page. That makes percentages very difficult. Has that been considered?
Jeanne: We haven't gotten to that level of detail yet. Once we've defined the user needs, we want to start writing the tests.
... So before we even write the guidelines, we want to write the tests.
... The tests should then show if the user needs have been met.
... For questions about quality, there would be manual tests that would be worth additional points. For things that could be automated, we would of course love automated tests.
Wilco: The work we do is not only about automation. It's about having a well-defined test procedure. The way that has been done is by drawing a line between "applicability" and "expectations". The applicability has to be purely objective, while the expectations can then be subjective. By doing that, you can then talk about percentages because you can enumerate the applicable content.
<Wilco> https://act-rules.github.io/rules/c4a8a4
Wilco: What would be the next steps for us to help out with this work?
Jeanne: I think I need to send you links to what we're working on. Then we can let you put in a link to your test, and we could then write the methods around it.
Wilco: We talked to AG and they were happy about us publishing the rules format.
<MoeKraft> I have to drop for another meeting.
Kasper: Currently 1 open pull request, which will spawn 2 more pull requests when merged. All need to be resolved before publishing.
Wilco: I agree, there's still work being done.
Kasper: We also need to account for the final call period if we want this to get done soon.
... Also, where is everyone?
... I will add a note to the rule about nested browsing contexts perhaps also needing default languages, but it might be something for a separate rule to check depending on how AT behaves. Wilco, could you verify how different AT treats missing default languages in iframes?
Wilco: Yes, I will do that.
<Wilco> https://github.com/w3c/wcag-act-rules/pull/12
Wilco: I have taken our rule, converted it to HTML, stripped out unneeded things, and opened a pull request. Will need to figure out how to display it.
... Will set up Netlify.
Wilco: I talked about this to a bunch of people, invited Jeanne and also talked to David MacDonald.
... Should we put together an email to send out?
... Do you want to draft something, Mary Jo?
Mary Jo: And send it to who?
Wilco: AG and WAI IG?
Mary Jo: Should we also reach out to members who aren't showing up? People are supposed to be contributing.
Wilco: I think so.
... Other business?
... See you in 2 weeks!