WebAIM - Web Accessibility In Mind

E-mail List Archives

Re: [EXT] Label in name

for

From: Sailesh Panchang
Date: Jun 25, 2021 2:51PM

Patrick comments: It's not that it does not apply. I'd rather say that
following H44 is also a good way to satisfy the SC.
Sailesh: That SC does not need to be tested when H44 is correctly
implemented. It increases the cost of testing / checking another box
needlessly.

Patrick comments: You'd still want to test to check that the
accessible name wasn't overridden on the control, negating the benefit
of H44...something like...
Sailesh:So one needs an SC to test that an accessibility technique
that has passed the tests for its correct implementation has not been
mauled by some other markup?

Patrick comments: Not really. It's more a side effect.
Sailesh: About the same side effect as an aria-label that is identical
to the visible text helps voice input users in addition to screen
reader user.

Patrick comments: The "ranks lower" part only comes into play if there
are multiple potential sources for the accessible name in play, as in
the above example.
Sailesh: The example is a very poor one. What purpose will one have in
using aria-label other than the intent to break correctly implemented
H44? If "Bar" is the label that makes sense for that control, then the
"Foo" label already fails SC 2.4.6 and several user groups.
The problem happens when instead of using techniques that rely on
native attributes / properties that are designed to help multiple user
groups gets overlooked because of implementer's penchant for using
WAI-ARIA without understanding the situations in which the use of
ARIA techniques is recommended.
e.g. using aria-labelledby in place of H44 for a 1-1 label-control association.

Patrick comments: Is there a point in having a "preferred" technique?
What matters is the end result, regardless of how the author achieved
it. There may well be valid situations where an author couldn't use
H44 directly, but still achieved the same result in another way. Which
is fine and passes.
Sailesh: I referred to situations when one cannot use H44 to make
content programmatically determinable or "extract and present
information to users in different modalities" . And I suggested those
techniques need to include guidance now contained in G208 / G211 which
is identical pretty much to SC 2.5.3. Accessibility consultants should
nudge developers to use the right technique for a particular
situation so that it helps the widest PWD user groups. Sometimes a
fallback technique may also need to be implemented to enhance
inclusivity without breaking other accessibility techniques.

Thanks,
Sailesh


On 6/23/21, Patrick H. Lauke < <EMAIL REMOVED> > wrote:
> On 23/06/2021 22:21, Sailesh Panchang wrote:
>> Some thoughts for consideration:
>> The Understanding doc states the SC does not apply when there is no
>> visible label.
>> In a sense it does not apply also when one has used technique H44
>> (for-id) method to associate the visible label with the control thereby
>> meeting SC 1.3.1.
>
> It's not that it does not apply. I'd rather say that following H44 is
> also a good way to satisfy the SC.
>
>> No further testing of SC 2.5.3 is required in this case.
>
> You'd still want to test to check that the accessible name wasn't
> overridden on the control, negating the benefit of H44...something like
>
> <label for="foo">Foo</label>
> <input id="foo" aria-label="Bar">
>
>> Therefore, even WCAG 2.0 considered the needs of voice input AT users;
>
> Not really. It's more a side effect.
>
>> Though the native HTML LABEL element (and thus H44) rank lower in the
>> accessible name computation
>
> The "ranks lower" part only comes into play if there are multiple
> potential sources for the accessible name in play, as in the above example.
>
>> it has benefits for multiple user groups and should be the preferred
>> technique.
>
> Is there a point in having a "preferred" technique? What matters is the
> end result, regardless of how the author achieved it. There may well be
> valid situations where an author couldn't use H44 directly, but still
> achieved the same result in another way. Which is fine and passes.
>
> P
> --
> Patrick H. Lauke
>
> https://www.splintered.co.uk/ | https://github.com/patrickhlauke
> https://flickr.com/photos/redux/ | https://www.deviantart.com/redux
> twitter: @patrick_h_lauke | skype: patrick_h_lauke
> > > > >