[css-overflow-5] Focus order of generated controls

[w3cbot]

This is a tracker issue. Only discuss things here if they are a11y group internal meta-discussions about the issue. Contribute to the actual discussion at the following link:

§ w3c/csswg-drafts#10912

[matatk]

APA discussed on 2024-12-11 and briefly recapped on 2024-12-18.

If we're reading into this specific question correctly, then there seems to be a deeper question about whether custom scrolling is going to work very differently from standard scrolling. With standard scrolling, you just have to be focused on, or in, the scrollable area, and the arrow keys will scroll the area. With this custom method, is the only way to scroll the area to focus and activate one of the buttons?

Have assigned myself and @AutoSponge for now to check this out; other APA volunteers welcomed!

[eric-hind]

Hi - have had a read through and had the following thoughts on this (which may be quite explainable)
Just to confirm, the spec is referring to rendered elements that will be part of the tab order?

1. How would these markers be represented to screenreader users - would it require ARIA investment per page or is there an HTML element that would make sense
2. How would this be supported for voice control - either through Dragon (web) or mobile voice control. Would the visual placement of labels or number markers be easy to see
3. In a zoom mode/scenario, would this add unnecessary scrolling for users who have optimized scrolling already?

[matatk]

This came up in a TAG design review (in fact it seems to touch on several). Notes about that can be found in today's APA minutes: https://www.w3.org/2025/02/26-apa-minutes#8672

In addition, on the related thread about CSS inert, @martinthomson raised the question of the related 'inert' property and how that would change how off-screen things would be accessed via the keyboard: w3ctag/design-reviews#1055 (comment) - the answer on an initial look seems ok if this is only going to be applied to carousels, but it also looks like some generalization of this scrolling model might happen in future, so we (APA) should track this and related developments to check on that.

APA should probably be tracking this via a meta-issue of some kind (cc @ruoxiran, @JaninaSajka). Also CC and thanks again @AutoSponge for your thoughts on the call today.