Update aria-details mapping for UIA

[travisleithead]

After reviewing the related issue w3c/html-aam#359, we determined that the UIA mapping for aria-detials needs an update.

@dlibby- reports that:

it should be mapped to Control Pattern: Annotation and Property: AnnotationObjects pointing at IDREFs of the value. We’d probably also want to put a note in about how the Annotation pattern exposes information.

There are also some textpattern specific mappings that I’m not sure fit cleanly in CoreAAM.

This bug tracks this intended update.

[scottaohara]

cc @benbeaudry

[benbeaudry]

Sorry for the delay.

I took the time to read the spec for aria-details, the UIA doc about the UIA_DescribedByPropertyId and the Annotation control pattern, I disagree with Daniel's suggestion of using the Annotation control pattern. In my opinion, UIA_DescribedByPropertyID is exactly what should be used for aria-details.

The UIA documentation defines the UIA_DescribedByPropertyId as

the DescribedBy property can offer quick access to the UI element that may already offer useful end-user information that describes the UI element.

This appears to match perfectly my understanding of the expectations the spec set for aria-details.

I discussed this offline with Daniel and he mentioned that he was under the impression that the UIA_DescribedByPropertyId did not allow ATs to jump to the linked elements. However, I quickly confirmed that this is possible with Narrator, a UIA-only screen reader, using the "Caps + A" command and the following minimal markup: data:text/html,<button aria-describedby=mybutton>hello</button><button id=mybutton>world</button>. Daniel now agrees that UIA_DescribedByPropertyId is a valid mapping for the aria-details attribute.

Since the core-aam spec already maps the aria-details attribute to the UIA_DescribedByPropertyId property, no further action is needed here.

[benbeaudry]

@scottaohara No changes are needed here, so closing this issue. Please reopen if you disagree. I continued to apply my reasoning from the previous comment and filed #204.