Add suggestion role

[aleventhal]

Relates to ARIA Annotations, issue #749

---

Preview | Diff

[joanmarie]

In addition to addressing the wording-related comments and questions, please also add suggestion to the appropriate list under https://w3c.github.io/aria/#roles_categorization

[carmacleod]

I think the words are ok, but I think this role needs an example to fully explain it.
I had to look at the explainer to understand where the markup would go (it's embedded in the content).
I think the -cat/+dog code example from the explainer would be helpful here.

[carmacleod]

Thanks for the example, @aleventhal!
Looks good!
+1

[joanmarie]

Looks good to me overall, but as I'm sitting here implementing support for this role in Orca, I have a question: Your text says:

The suggested change must be present as child content, as an insertion, a deletion, or one of each.

Can there be other children, like a paragraph explaining why the proposed insertion and/or deletion should be made? Or a time stamp? Or am I guaranteed to only find insertion and/or deletion children?

Arguably Orca should handle both scenarios because authors. But it would be nice to know (and perhaps see text in the spec stating) what the right usage looks like. Also, if we do want to prevent authors from putting non-insertion/deletion children in a suggestion, adding this statement to the spec should then cause it to be checked by validators.

[aleventhal]

@joanmarie PTAL, I updated the text to make this more clear.

Looks good to me overall, but as I'm sitting here implementing support for this role in Orca, I have a question: Your text says:

The suggested change must be present as child content, as an insertion, a deletion, or one of each.

Can there be other children, like a paragraph explaining why the proposed insertion and/or deletion should be made? Or a time stamp? Or am I guaranteed to only find insertion and/or deletion children?

Arguably Orca should handle both scenarios because authors. But it would be nice to know (and perhaps see text in the spec stating) what the right usage looks like. Also, if we do want to prevent authors from putting non-insertion/deletion children in a suggestion, adding this statement to the spec should then cause it to be checked by validators.

[joanmarie]

I like it! Thanks for adding clarity regarding other possible children!!

[mcking65]

No big concerns, but I think it is very important that the normative language be specific, especially wrt the responsible party. Most of my suggestions are related to that concern.

[aleventhal]

@mcking65 I think I got everything, but I needed to do some polish after as there was an incomplete sentence snippet that remained. Can you take a look?

[mcking65]

Looks good. I just made one minor editorial suggestion.