Meeting minutes
<Lionel> https://
Bev: I work in various groups at W3C, NIST, et al, retired from employment...
Lionel: I am COO of ?? - have been working in accessibility for a while, on automation inter alia, and going to see Wagner in Feb (I'm a musician too)
Chaals: I've been working in standards for quarter-century, was working for WAI a long time ago, currently working as lead Standards Architect at ConsenSys (Ethereum company)
… and seconded as Technical Program Director to Enterprise Ethereum Alliance.
Chaals: Suggest using a github issue that asks people to leave an introduction as a comment on the issue.
<Lionel> https://
[general agreement]
intro to what we're doing
JS: started with the controversy over overlays - they cna be problematic but they can also be useful tools. But the scope is broader than that, about stuff done at the edge in general.
… APA group had some good discussions with groups at W3C TPAC event...
… might want to use secure privacy-preserving credentials to set up things we used to try to do with Stylesheets - e.g. carry settings across devices...
… or what happens when you need a legal document, and want that to be in accessible form.
… but we started with overlays, looking at a definition of them and identifying anti-patterns in current deployment. Think we should distinguish those that work on the fly vs those that are user-directed.
… Would like to move our draft forward...
chaals: We don't need so much of a definition of it, more 'what does it do'
… I can think of various overlays, also from other domains
… it's important to identify patterns as well as anti-patterns
… to help people reason about which parts they want, and which they don't...
… and even find ways to identify virtuous patterns, and even integrate parts of various overlay behaviors together
Janina: There may be more than one thing that has overlaid the rendering, and a conflict resolution might be valuable
Chaals: Yes, that would be nice.
Janina: That's a goal. How do we get there - let's define good behaviour as well as bad behaviour
Lionel: I am happy to go without a really strong definition (although I like having one). We've heard that idea before...
… focus on the outcomes might be helpful.
… Agree that a catalogue of what it does would be helpful. Then we get into things like descriptions we can build on.
Janina: Action seems to be adding a section for beneficial applications.
… under development
Lionel: "Virtuous Patterns" that add outcomes that are desirable, and "Anti-Patterns" doing things that are not desirable.
Janina: Things that you can't turn off might be even worse than an anti-pattern, because there is no way to control it.
chaals: If we label 'good' pattern and 'bad' pattern, we might be in for a lot of argument
… start with, he's what overlays do. What is the goal and the eventual impact.
… Proposing the header, "What Overlays Do"
Lionel: Like it better than trying to call things good and bad ... value language leads to difficulties.
chaals: At some point we will have value-laden language, as the key is the measurable impact on users
… if the measurable impact is a positive, then that's good. If the impact is not good, that's not a win.
Janina: Any technology that can be useful can be a problem. So we get to "what is an overlay trying to do, how does that work out, what makes the difference?"
Lionel: We do need to get beyond the "this technology is all bad" simplifications to point out the useful bits (and guide people away from the bad bits)
Chaals: Is there a list of overlays we can look at...
Lionel: Karl Groves made one...
Janina: Make sure we have permission to use whatever we are using.
Lionel: Do we ask them? Do we look at the websites?
chaals: Start by everyone taking one, and collect a (messy) list, this does this, and this, and this
… then once we start getting the information together
… we can look for ways to group it
an issue describing a list of tools to review
Chaals: I'll take on "Amaze" - see https://
Lionel: I'll take on my own tool first :)
Next meeting
Janina: let's not try boxing day or January 2 ...
RESOLUTION: Next meeting 9 January