ARIA required owned, do we need the no implicit requirement? bc4a75

[WilcoFiers]

Got some more feedback on the following line, this time from the ACT TF:

the element has an implicit semantic role that is identical to its explicit semantic role; or

Now that I've seen what the content model rules look like, this might not be necessary anymore. Need to think through the different cases.

[Jym77]

I must say I was never a big proponent of the "no implicit" exception in "required context role" and "required owned element" rules…

As far as I remember, this exception is due to the fact that with implicit role the lack of required context/owned would also be a failure of 4.1.1 Parsing. I think the situation is very similar to the "button has accessible name" rule carefully avoiding image buttons because they would also fail 1.1.1.

We kinda discussed the situation (for button/image buttons) and concluded (iirc) to remove the "image button" exception (even though this was never really finished @EmmaJP ) See #1049, #1184 and https://www.w3.org/2020/02/13-act-r-minutes.html#x03

I think we are in a similar situation here. My position would be to remove the "no implicit" exception which is very weird and forces the examples to also be very weird (notably breaking the First rule of ARIA).
I am also not sure how frequent it is "in the wild" to have an explicit role different from the implicit, and with required context/owned roles…

Note that in Alfa we decided to ditch the "no implicit" exception from both context/owned rules as it streamlines things a lot.

[ajanec01]

Hey @WilcoFiers

Would it be worth adding it as an agenda item? I'm happy to create a PR to clear issues related to the feedback on ARIA required owned elements, but I think it would be nice to have some discussion on it beforehand.