Re: svg accessibility tests

Hey Fred.

On 03/04/2016 11:13 AM, Fred Esch wrote:
> Joanie,
> That calculation is done based on the core AAC below, emphasis is mine.
> 4.3 Text Alternative Computation
> <>
> Text alternatives are built up, when appropriate, from all the relevant
> content contained within an element. This is accomplished via steps 2B
> and 2F, which are recursive, *using the full set of rules*to retrieve
> text from its own children or nodes it references.

The full set of rules also include a statement that you only do the
processing if "the current node is not already a part of an
{aria-labelledby,aria-describedby} traversal. See issue-1008.


> Reasoning is inline below. If my interpretation is wrong, please let me
> know. If it is unclear what should happen please mention this and we can
> bring it up as an agenda item.

I think your interpretation may indeed be wrong. Again, see issue-1008.


>> In the meantime, please refresh my memory regarding what happened with
>> the recursive labelledby/describedby. Shouldn't this test:
>> // two valid idrefs, one with an aria-label, one with aria-labelledby
>> if given
>>  <circle id="test" aria-labelledby="t1 t2" cx='50' cy='100' r='15'/>
>>  <text id="t1" x="40" y="90"aria-label="big">the</text>
>>  <text id="t2" x="140" y="90" aria-labelledby='t3'>end</text>
>>  <text id="t3"x="180" y="90">finish</text>
>> then accessible name = "big finish"
>> have accessible name of "big end"?
> *the element with id t1 would contribute **/big/** *


> *the element with id t2 would contribute **/finish/** because of the
> aria-labeledby.*


Looking at

This part is true:

    if computing a name, and the current node has a non-empty
    aria-labelledby attribute

However, this part is false:

    and the current node is not already part of an aria-labelledby

The current node (t2) IS already part of an aria-labelledby traversal,
namely the node with id="test" with aria-labelledby="t1 t2".

Because of the above condition failing, we cannot do this bit:

    process its IDREFs in the order they occur

Instead, we wind up having to use the text, namely "end".


Received on Friday, 4 March 2016 17:03:12 UTC