Explore concept/styling for direct advice for how AT/screen readers `should` interpret certain aria roles/values

[spectranaut]

At the ARIA F2F we discussed how to better our relationship with screen reader developers: https://www.w3.org/2023/05/04-aria-minutes#t05

One thing that came out of the discussion was that some ATs would appreciate guidance on how to implement support for new ARIA features. We will explore a process for drafting AT guidance directly in the spec under the feature and role. As part of the consensus process we can get feedback from ATs through our many personal relationships.

Additionally it seemed like there was some consensus to use "should" and "may" with respect to ATs, and acknowledged that often we mean screen readers specifically.

This issue is for the editor team to explore the styling of these sections and to provide draft guidance for a recent feature, for example, the annotation roles.

[spectranaut]

a topic for a future editors meeting, @pkra

[stevefaulkner]

Related to this: I recently did some testing on roles with presentational children What I found was that Browsers interpreted the requirements differently and SRs dealt with this sub optimally in some cases. It would be helpful for operability and user experience if in the case of User Agents SHOULD or SHOULD NOT, that some clear guidance be provided.

[pkra]

We discussed this on the editors' call yesterday https://lists.w3.org/Archives/Public/public-aria-editors/2023Jun/0002.html

One decision was to resolve this issue here by adding editorial advice to https://github.com/w3c/aria#editorial-documentation

The issue on aria-common https://github.com/w3c/aria-common/issues/96 is meant to solve the design side of things.

[pkra]

As per ARIA Editors meeting 2023-06-26, we want to start by prototyping this here, limited to a specific section.

Labeled Agenda to decide on a suitable section and someone to take this on.

[jnurthen]

I will try to prototype something in the next month.