Request for aria-completed state

[brennanyoung]

In several of our e-learning projects, we have needed to indicate the completion state of a section/lesson. We do this visually with a checkmark on our links and 'progress overview' pages, but can't find anything in ARIA which would express that something is complete.

The checkbox/radiobutton semantics come closest (because boolean), but those are intended for operable roles, and I am looking for something read-only. The only other thing I can imagine using is aria-busy on a progress bar role, but that doesn't seem great either.

I've been advised to concatenate the word "complete" or "incomplete" onto the aria-label, which strikes me as a hack, (although it works). Plus, it adds a new arbitrary string to our translation catalog, which I would prefer to have handled by the AT.

[brennanyoung]

Just discovered that this is similar to #1000

[scottaohara]

@brennanyoung is this issue a duplicate of the previous issue then?

[brennanyoung]

@scottaohara not exactly, it is more like an additional and separate use case which suggests the need for such an attribute may be broader than the case mentioned in issue #1000 - it doesn't mention "completed" (so doesn't appear in a search for that term), but instead calls for an opposite to "invalid" - i.e. "this input is flagged as good".

"Completed" or "done" or "not invalid" might well express the same thing (the exact name of the attribute could be derived from any of these or something similar), but could be used in places other than form inputs.

With a separate issue, we can first discuss whether 'not invalid' and 'completed' are the same thing. If so, we can close one of the issues.

[brennanyoung]

Just revisiting this issue for a bump because we are implementing a progress read-only breadcrumb widget for a quiz/test. Mockup looks like this:

Note: The orange boxes are not intended to be part of the GUI, rather provided as guidance for the devs.

So here, we are obliged to use img or role=img for the circles, otherwise there will be no aria-label or alt attribute to append the string "complete".

But semantically, these are not images, they are list items. AFAICT screen reader support for aria-label is not to be expected for the listitem semantic role.

Of course we can insert an additional element inside the list items, and give it an image role, but shoehorning the steps in to an image role just so that we can append a string to express "complete" (or "incomplete") seems very hacky to me. A proper aria-completed would allow content which does not support aria-labels to express this.

So, in answer to @scottaohara : The need to express 'complete/incomplete' on non-operable content does indeed seem distinct from the use case mentioned in issue #1000 (which is about operable/form elements), although aria-completed might still serve both use cases.

[aleventhal]

Another option is aria-describedby, and the new aria-description.
IMO, avoiding a localization is not a good enough reason for a new property. There are already too many ARIA properties. A good reason would be if the AT needs to know about it because it alters the interaction model, and if it would solve an extremely common use case that can't be handled satisfactorily in another manner.

[brennanyoung]

OK, the current draft spec for aria-description seems to handle this case.

In particular:

In cases where providing a visible description is not the desired user experience, authors MAY set the accessible description of the element using aria-description.

...and for "all elements of the base markup" covers it, including non-operable roles.

As long as both of these conditions apply, then we can close this. Thanks.

[carmacleod]

@brennanyoung

Given that aria-description is brand-new and not really supported yet, note that aria-describedby can be used to point to hidden content: https://www.scottohara.me/blog/2017/04/14/inclusively-hidden.html#exploiting-a-completely-hidden-content-loophole

Also, aria-current="step" might be a nice addition: https://w3c.github.io/aria/#aria-current
It is well supported in screen readers: http://design-patterns.tink.uk/aria-current/