Seeing many blank aria-labelledby attributes in BASE

[terracoda]

@jessegreenberg, @zepumph, @mbarlow12, I am seeing blank aria-labelledby="" attributes on all elements in BASE. I'm not sure why these attributes have appeared en masse.

Are you guys working on something?
Is something in transition?
Perhaps, there is an open issue in another repo about this?

[terracoda]

Screen shot from BASE on Master

[terracoda]

Has anyone checked to see if null strings are allowed (valid code) for aria-labelledby?

[jessegreenberg]

Hmm probably caused by changes in phetsims/scenery#701

[jessegreenberg]

Thanks for catching this @terracoda, should be fixed in the above commit, those blank aria-labelledby attributes have been removed.

[terracoda]

I am totally not up to speed on what you guys are implementing, but I am wondering if you considered options for creating an accessible name? Perhaps this is too high-level.

• option 1: inner text of element, e.g.,

<button>My Button's Name</button>

• option 2: using for attribute on a label element to associate an html label with html control element, e.g.,

<input id="radiobuttonIDREF" type="radio"><label for="radiobuttonIDREF">My Radio Button's Name</label>

• option 3: using aria-label attribute to label an html element that does not have inner text, or label or other way to provide an accessible name, e.g.,

<input id="checkboxIDREF" type="checkbox" aria-label="My Checkbox's Name">

• option 4: using aria-labelledby attribute to label an html element that can be associated with an element in the PDOM, like a heading element that would provide the elements name inner text, e.g.,

<section id="sectionIDREF" aria-labelledby="headingIDREF">
<h2 id="headingIDREF">My Section's Name</h2>
....
</section> 

Of course, examples are simplified, and I am not trying to imply an order of priority with these options. There may be diverse reasons (interaction preferences, pedagogical reasons, technical reasons, visual design) on which naming option to choose.

[terracoda]

Thanks @jessegreenberg, I'll close again as the issue has been resolved, but leave my comment.