Matthew: i'm researching - we discussed at TPAC - a fundamental question is : we did talk about this in CSS but the minutes haven't appeared yet.

Yves: finds minutes and minutes

Matthew: will look more into it and see if question I had was answered... A question I had around the whole conversation of reading order. Talking about focus navigation and not how the DOM order is exposed - virtual cursor navigation. That follows the DOM order. There's been discussion on whether that needs to change as well.

... tables or things that look like tables...

... the current stuff, good progress is being made...

Matthew: Haven't reviewed this yet.

Ditto

Matthew: a few thoughts... some general issues - and specific a11y discussions. I haven't seen anything in this proposal that's concerning. But a couple of things - reading the spec from the CSS WHATWG side... About reading flow in general ... Specific words & phrases might have an a11y implication - e.g. sequential navigation order is clear, but "speech order" is being referred to in the spec... https://drafts.csswg.org/css-display-4/#order-accessibility CSS WG is capable... But APA and CSS might need to get together to understand these common terms... So no specific concerns about this review but I think APA needs to talk to CSS... Action on me. ... when I read the spec it's more focused on the order of interactive elements... could supersede tab index.

Jeffrey: my sense is that this should affect the a11y tree. No definition of how html turns into an a11y tree...

Matthew: one interesting thing - talks about the difference between source order and display order. If it does effect a11y tree then why not just reorder source?

Tess: The a11y tree comes from the render tree. But different ATs are different.

some discussion on this point

Peter: would be good to have a reference in the CSS spec.

Matthew: also the explainer is in a personal repo.

Jeffrey: Will poke. It should be in "explainers by googlers"...

Matthew: "render to speech" different from the "focus order".

Jeffrey: it's the wrong term, regardless... Seems like a good issue to file with CSS.

Peter: just go ahead and post your points to the public thread.

Matthew: as expected - the thing I was checking for was confirmed - leads to another question... it does effect everything - from the a11y tree it will impact the visual order. The question next : whether this means - e.g. if you have masonry layout if there is a change in the viewport size and the layout gets changed, should we need events to AT to let it know... Usually don't need this but I don't know. I could ask an AT vendor... perhaps we should ask this question about whether assistive technologies had got far enough down the path to know this... Given we've got the 2 people in this thread - I'll post the question and then redirect to an APA thread if so...

Dan: just post that in our thread then...

Matthew: Will post comment soon (before plenary)

Matthew: I need to reply with a comment...

Peter: we were happy to close it last week... had your question on the a11y order...

Matthew: we were wondering if they were aware of / work with AT vendors... APA is asking our contacts... Will post later today and then we can close it.

Peter: is there a defined eventing system for a11y tools that aren't js?

Matthew: it's outside of the viewport... There are several different APIs that these tools use to talk to the browser. It's something we're (APA) looking into separately. From the spec perspective it maybe shouldn't be in the spec - or it might be an editor's note.

Peter: there should be a spec for all of that - standardized behaviour across all browsers.