RE: ARIA annotations open issues

On May 9, 2014, at 12:40 AM, Marco Zehe <marco.zehe@googlemail.com> wrote:

> Am 08.05.2014 23:12, schrieb James Craig:
>>> Action to load annotations
>>> There should be a way for the screen reader to ask the website to load the annotation if hidden through some new protocol. IndieUI is one possible option here.
>> 
>> Marco, IIRC, this was your request. Will you please remind us why you think it's critical that the screen reader user have access to information that is not available in the mainstream interface?
> 
> This was in reference to Word for Windows where a sighted user could just hover the mouse over the comment indicator and see the comment text pop up in a tool tip, without opening the comments pane. JAWS and other screen readers provided the same info as soon as the regular editing cursor landed on the comment or footnote or whatever indicator, to give a blind user the same level of access without requiring them to open the pane.
> 
> So if in Word for the web, the sighted user also has to perform an action to actually make the comment visible like clicking an option in a context menu rather than just hovering the mouse over the annotation indicator, this point can be dropped. I was under the impression that a similar mechanism like in Word for Windows would work here, but if it isn't, this point is not necessary.

Okay. This is already achievable via :focus/onfocus then. The hover and focus events and pseduo-classes would display ARIA tooltips for hover states, or the full annotation connection if the comment window was displayed. We could take an action for the authoring practices guide.

Chris, does the existing ARIA tooltip pattern suffice for your needs on the hover/focus example Marco mentioned? 

[Chris] I think that overall, the problem of lazy loading is something that's not specific to annotations, so we probably shouldn't create an action that's specific to annotations. Practically speaking, I'm not sure that adding a tooltip (custom or standard) to the text would work since very few screen readers support reading out anything within a contenteditable when the user's arrowing around, let alone a span that has a title attribute that contains the comment text.

Received on Friday, 9 May 2014 20:05:24 UTC