WebAIM - Web Accessibility In Mind

E-mail List Archives

Thread: Does filtering count as "change of context" as regards WCAG 3.2.2

for

Number of posts in this thread: 4 (In chronological order)

From: Robert Fentress
Date: Fri, Jun 16 2017 10:46AM
Subject: Does filtering count as "change of context" as regards WCAG 3.2.2
No previous message | Next message →

A pattern you see more and more is when a user changes the setting of an
interface component in a web page, such as a select element, checkbox or
text field, and it dynamically filters the results displayed in the main
content area of the page. Does WCAG 3.2.2
<https://www.w3.org/TR/UNDERSTANDING-WCAG20/consistent-behavior-unpredictable-change.html#context-changedef>
require, then, that the user be informed that making such changes will have
this effect before he or she begins interacting with the control? That
criteria as is as follows, by the way:

*On Input:* Changing the setting of any user interface component does not
automatically cause a change of context unless the user has been advised of
the behavior before using the component.


If the user must be informed, is it sufficient to make a blanket statement
somewhere on the page before these controls that says something like
"changing the value of these controls dynamically filters the results
below." For web applications, would it be sufficient to provide
instructions in the documentation. Or must the control be semantically
associated with the in-page instructions, by doing something like adding
aria-describedby to the control and pointing to the id of the
instructions? Would doing so make things too verbose? My justification
for making that explicit association is that users may just tab to the
control and miss the instructions, otherwise.

Also, I like the idea of using a polite ARIA live region to announce the
number of results returned in a non-verbose way, like "5 results." Anybody
think that is bad practice? Just seems like that is a useful affordance
that more closely approximates the experience of sighted users. What do
you think?

Best,
Rob

--
Rob Fentress
Senior Accessibility Solutions Designer
Assistive Technologies at Virginia Tech
Electronic Business Card (vCard)
<http://search.vt.edu/search/person.vcf?person54847>
LinkedIn Profile
<https://www.linkedin.com/in/rob-fentress-aa0b609?trk=profile-badge>

From: Tim Harshbarger
Date: Fri, Jun 16 2017 11:24AM
Subject: Re: Does filtering count as "change of context" as regardsWCAG 3.2.2
← Previous message | Next message →

I think the key in this situation is whether or not the change in content is a change in context.

When you read further on this SC, it talks about context. It also states that a change of contents on the page may or may not represent a change of context.

My own opinion is that applying the filtering options does not change the context. However, adding a live region to report on how many results remain after applying a filter would be a really great idea.

I know I would find it useful to receive immediate feedback regarding whether or not the filter I just applied left me with 5 items or 0 items. When that feature isn't present, it doesn't necessarily prevent me from using filtering or using the page, but it means I do have to try to figure out what results remain. Being told immediately how many results remain would make it easier to know if I had applied too many filters or not enough.

From: Matt King
Date: Fri, Jun 16 2017 1:15PM
Subject: Re: Does filtering count as "change of context" asregardsWCAG 3.2.2
← Previous message | Next message →

I completely agree with Tim.

From: Birkir R. Gunnarsson
Date: Sat, Jun 17 2017 5:45AM
Subject: Re: Does filtering count as "change of context" as regards WCAG 3.2.2
← Previous message | No next message

Absolutely. The only thing I add to this is establishing an
aria-controls relationship between the controling element (the
dropdown or checkbox or what have you) and the element containing the
data that is affected (e.g. qa table).
Admittedly it doesn't do much, and support is not grade, but I believe
it is semantically good information to add, at the cost of one
attribute (2 if you need to add an id attribute on the controlled
element just for that purpose).


On 6/16/17, Matt King < = EMAIL ADDRESS REMOVED = > wrote:
> I completely agree with Tim.
>
>