888256 - [AccessFu] Speak landmarks

Status: RESOLVED FIXED

(Core :: Disability Access APIs, defect)

Description

[Marco Zehe (:MarcoZ)]

Landmarks are a newer accessibility feature of HTML. They are defined by either using role="<landmark_role>" or by using the HTML5 equivalent element. <landmark_role> is one of banner, complementary, contentinfo, main, and navigation. the landmark info is saved in the xml-roles object attribute. It may be on its own, or coupled with other roles, since the ARIA role attribute can have more than one value at one given time. So the search is a "contains", not an "is".

To implement,

a) When traversing from element to element, if the accessible or one of its parents has a landmark attribute, and that differs from the previous, speak a phrase like "in banner" or "in main". Note that other screen readers also speak the landmark that was just exited again, so that you hear something like "exiting banner, entering main", but I think that is too excessive. Stop at the first landmark found, I don't think supporting nested landmarks is necessary.

b) Add the ability to use quick navigation gestures to select landmark as a type, and also add a key to go to the next or previous landmark. NVDA uses the letter d, JAWS uses the semicolon. I think sticking with d as the NVDA equivalent is a good idea.

c) In braille, just put the word of the actual landmark, no abbreviation or full role or other word like "in".

d) Add utterance and braille tests to make sure they are accounted for.

An example of landmark tests for the API can be found here: http://mxr.mozilla.org/mozilla-central/source/accessible/tests/mochitest/attributes/test_xml-roles.html?force=1

Comment 1

[Yura Zenevich [:yzen] (please use "needinfo")]

Attachment: Patch for 888256 v1.

Just a WIP to get some feedback, tests are still to come.

Marco, I was not 100% certain how the correct output for braille should look like in cases when we also add a '*' or an accessible's previous sibling name to the beginning of the braille output.

Comment 2

[Eitan Isaacson [:eeejay]]

Comment on attachment 771615 [details] [diff] [review]
Patch for 888256 v1.

Review of attachment 771615 [details] [diff] [review]:
-----------------------------------------------------------------

This looks good.

::: accessible/src/jsat/OutputGenerator.jsm
@@ +164,4 @@
>      }
>  
>      if (name) {
> +      let outputOrder = typeof gUtteranceOrder.value === 'number' ?

Not entirely necessary/relevant, but I could see how this helps with consistency.

@@ +183,5 @@
> +      return;
> +    }
> +
> +    let outputOrder = typeof gUtteranceOrder.value === 'number' ?
> +      gUtteranceOrder.value : this.defaultOutputOrder;

Instead of doing this again and again, why not have in the global scope:

function outputOrder() {
  return typeof gUtteranceOrder.value === 'number' ?
		gUtteranceOrder.value : this.defaultOutputOrder;
}

Comment 3

[Yura Zenevich [:yzen] (please use "needinfo")]

Comment on attachment 771615 [details] [diff] [review]
Patch for 888256 v1.

Review of attachment 771615 [details] [diff] [review]:
-----------------------------------------------------------------

::: accessible/src/jsat/Utils.jsm
@@ +203,2 @@
>  
> +    if (!this._attributes) {

Need to get rid of caching.

Comment 4

[Marco Zehe (:MarcoZ)]

Comment on attachment 771615 [details] [diff] [review]
Patch for 888256 v1.

f=me with the changes we discussed f2f w.r.t form and application, removal of "in " etc.

Comment 5

[Yura Zenevich [:yzen] (please use "needinfo")]

Attachment: Patch for 888256 v2.

Comment 6

[Marco Zehe (:MarcoZ)]

Comment on attachment 772351 [details] [diff] [review]
Patch for 888256 v2.

r=me with one nit: Can you correct all instances of "an section" to read "a section", reads funny otherwise. ;-)

Comment 7

[Yura Zenevich [:yzen] (please use "needinfo")]

Attachment: Adding support for speaking landmarks. r=eeejay, marcoz

Comment 8

[Eitan Isaacson [:eeejay]]

Comment on attachment 772700 [details] [diff] [review]
Adding support for speaking landmarks. r=eeejay, marcoz

Review of attachment 772700 [details] [diff] [review]:
-----------------------------------------------------------------

This looks good. Just a XXX comment near the brailled list item, like we said.

Comment 9

[Yura Zenevich [:yzen] (please use "needinfo")]

https://hg.mozilla.org/integration/mozilla-inbound/rev/24c5584de220

Comment 10

[Ryan VanderMeulen [:RyanVM]]

https://hg.mozilla.org/mozilla-central/rev/24c5584de220