WebAIM - Web Accessibility In Mind

E-mail List Archives

Re: Hidden headings with aria-labelledby or just aria-label best for for labelling landmark regions?

for

From: Steve Green
Date: Jul 23, 2018 1:02PM

There is a surprisingly difficult balance to be struck when acting on the results of user testing. We aim to:

1. Include features we know will benefit the user groups we tested with.

2. Make sure we don't remove features that might benefit the user groups or people of proficiency levels we didn't test with. As a third party, we don't control the testing budget, so we never get to test with as many people as we would like.

3. Include (or at least not remove) features that might benefit people in the future when user agent support and awareness of accessibility features improve.

4. Not include so many of these features that we create a terrible user experience for some people now.

5. Be aware that proficiency levels vary more than they do for non-disabled people, and take account of the different strategies that people of different proficiency levels use for navigating in and between pages.

There's probably more, but that's plenty to be thinking about.

Steve


-----Original Message-----
From: WebAIM-Forum < <EMAIL REMOVED> > On Behalf Of glen walker
Sent: 23 July 2018 19:39
To: WebAIM Discussion List < <EMAIL REMOVED> >
Subject: Re: [WebAIM] Hidden headings with aria-labelledby or just aria-label best for for labelling landmark regions?

Good stuff, Steve, and thank you for caveating your results of landmarks
(#4) by saying that your results don't mean we shouldn't use them.

The same is true for skip links (#1), although your results were qualified by screen reader users. As a keyboard user, I use skip links quite often.


On Mon, Jul 23, 2018 at 11:40 AM, Steve Green < <EMAIL REMOVED> > wrote:

> Of the participants who have expressed an opinion, every one of them
> expects the <h1> to be at the top of the main content - they really
> don't care if it is preceded by lower heading levels. Behaviours I
> have observed over 14 years of user testing include:
>
> 1. Very few screen reader users use skip links even when they are
> provided and work correctly. Most people never even hear them because
> the screen reader goes into SayAll mode when a page loads, and they
> just stop it and start navigating from wherever it stopped.
>
> 2. Very few screen reader users use the number keys to jump directly
> to an <h1>. Almost all press the H key repeatedly till they hear the
> <h1>. This is only slightly less efficient and it's slightly easier to
> find the H key than the 1 key.
>
> 3. Consistency of structure within a page and between pages is far
> more important than perfect nesting of headings.
>
> 4. I have not seen anyone navigate by landmarks. With the exception of
> "obvious" landmarks such as main and navigation, most people don't
> know what the landmarks are for. As such, they are regarded as noise.
> That's not to say we shouldn't use them, but at the moment they are
> not understood well enough to be useful.
>
> Steve
>