E-mail List Archives

Search E-mail List Archives for

Number of posts in this thread: 14 (In chronological order)

From: Michael Ausbun
Date: Tue, Sep 17 2019 9:16AM
Subject: 1.4.3 & 2.4.7: evaluation while blind
No previous message | Next message →

Greetings,
I am looking for a more efficient and consistent way to evaluate for 1.4.3 contrast and 2.4.7 focus visible. As a blind person, what might be (in your opinion) the most efficient way of evaluating for these?
I appreciate your thoughts!
Respectfully,
Michael

From: Jared Smith
Date: Tue, Sep 17 2019 9:53AM
Subject: Re: 1.4.3 & 2.4.7: evaluation while blind
← Previous message | Next message →

We've heard from blind users that indicate WAVE (wave.webaim.org)
works well to evaluate poor contrast text. I think focus visible would
be a bit difficult because it is typically cannot be accurately
evaluated in an automated way.

Jared Smith
WebAIM.org

From: mhysnm1964
Date: Wed, Sep 18 2019 12:34AM
Subject: Re: 1.4.3 & 2.4.7: evaluation while blind
← Previous message | Next message →

Axe that is a plugin for Firefox and is built into chrome also can provide
this information.

From: Birkir R. Gunnarsson
Date: Wed, Sep 18 2019 5:07AM
Subject: Re: 1.4.3 & 2.4.7: evaluation while blind
← Previous message | Next message →

Unless reported by automated scans, these are two success criteria I
do not try to evaluate, instead I have a sighted tester handle it.
Keyboard focus is easy to test with a pair of eyes but really hard to
test with non-working ones.
There is an NVDA plugin that tries to indicate whether focus is
visible but I've never used it.

Color contrast can be complex and somewhat subjective if not simple
enough to be detected by a scanner such as Wave or Axe (e.g.
gradients, charts).


On 9/18/19, = EMAIL ADDRESS REMOVED = < = EMAIL ADDRESS REMOVED = > wrote:
> Axe that is a plugin for Firefox and is built into chrome also can provide
> this information.
>
>

From: Patrick H. Lauke
Date: Wed, Sep 18 2019 5:10AM
Subject: Re: 1.4.3 & 2.4.7: evaluation while blind
← Previous message | Next message →

On the same topic, I suspect 1.4.10 Reflow, 1.4.11 Non-Text Contrast and
1.4.12 Text Spacing may also prove challenging. As might 4.1.3 Status
Messages (due to the very nature of the failure itself having things not
announced by AT, and potentially disappearing completely from the DOM
before an AT user can go through the page to verify if any such status
was displayed but not announced).

P
--
Patrick H. Lauke

www.splintered.co.uk | https://github.com/patrickhlauke
http://flickr.com/photos/redux/ | http://redux.deviantart.com
twitter: @patrick_h_lauke | skype: patrick_h_lauke

From: Steve Green
Date: Wed, Sep 18 2019 5:30AM
Subject: Re: 1.4.3 & 2.4.7: evaluation while blind
← Previous message | Next message →

I don't see how a blind person can test several others, such as 1.3.1 Info and Relationships. That SC requires you to assess the visual presentation and check if the information and relationships are conveyed programmatically. How can you do that if you can't see it?

There are others like 1.3.2 Meaningful Sequence. If you can only perceive the programmatic reading sequence and not the visual sequence, how can you tell if it is correct?

And 1.4.1 Use of Colour. How can you tell if colour alone is being used? It might be possible by analysing the styles, but that won't work if the colour is in images.

I won't labour the point, but if you can't perceive something, how can you know if there is anything to perceive?

Steve Green
Managing Director
Test Partners Ltd

From: Patrick H. Lauke
Date: Wed, Sep 18 2019 6:12AM
Subject: Re: 1.4.3 & 2.4.7: evaluation while blind
← Previous message | Next message →

On 18/09/2019 12:30, Steve Green wrote:
> I don't see how a blind person can test several others, such as 1.3.1 Info and Relationships. That SC requires you to assess the visual presentation and check if the information and relationships are conveyed programmatically. How can you do that if you can't see it?
>
> There are others like 1.3.2 Meaningful Sequence. If you can only perceive the programmatic reading sequence and not the visual sequence, how can you tell if it is correct?
>
> And 1.4.1 Use of Colour. How can you tell if colour alone is being used? It might be possible by analysing the styles, but that won't work if the colour is in images.
>
> I won't labour the point, but if you can't perceive something, how can you know if there is anything to perceive?

Oh, absolutely. Sorry, wasn't trying to make an exhaustive list :)

P
--
Patrick H. Lauke

www.splintered.co.uk | https://github.com/patrickhlauke
http://flickr.com/photos/redux/ | http://redux.deviantart.com
twitter: @patrick_h_lauke | skype: patrick_h_lauke

From: Birkir R. Gunnarsson
Date: Wed, Sep 18 2019 6:22AM
Subject: Re: 1.4.3 & 2.4.7: evaluation while blind
← Previous message | Next message →

1.3.1 is typically fairly perceivable. A line of text that marks the
start of a section should be a heading, If you read the page in
context you can typically detect what the heading structure should be.
you can detect table markup issues fairly easily just from the content,
1.3.2 is all about logical content order so it's more easily
detectable using a screen reader in fact (the requirement is that it
is logical, not necessarily that it matches visual order, though
preferred).
1.4.1 is tricky but you can typically see that from class names, and
you can use a Jaws scripts to display class names for any element you
can reach in browse mode, not just keyboard focusable). I still always
verify that with a pair of eyes.
I should write an article on this, been meaning to for ages.




On 9/18/19, Patrick H. Lauke < = EMAIL ADDRESS REMOVED = > wrote:
> On 18/09/2019 12:30, Steve Green wrote:
>> I don't see how a blind person can test several others, such as 1.3.1 Info
>> and Relationships. That SC requires you to assess the visual presentation
>> and check if the information and relationships are conveyed
>> programmatically. How can you do that if you can't see it?
>>
>> There are others like 1.3.2 Meaningful Sequence. If you can only perceive
>> the programmatic reading sequence and not the visual sequence, how can you
>> tell if it is correct?
>>
>> And 1.4.1 Use of Colour. How can you tell if colour alone is being used?
>> It might be possible by analysing the styles, but that won't work if the
>> colour is in images.
>>
>> I won't labour the point, but if you can't perceive something, how can you
>> know if there is anything to perceive?
>
> Oh, absolutely. Sorry, wasn't trying to make an exhaustive list :)
>
> P
> --
> Patrick H. Lauke
>
> www.splintered.co.uk | https://github.com/patrickhlauke
> http://flickr.com/photos/redux/ | http://redux.deviantart.com
> twitter: @patrick_h_lauke | skype: patrick_h_lauke
> > > > >


--
Work hard. Have fun. Make history.

From: Detlev Fischer
Date: Wed, Sep 18 2019 6:31AM
Subject: Re: 1.4.3 & 2.4.7: evaluation while blind
← Previous message | Next message →

In our experience, there are many aspects of 1.3.1 that CAN be tested by
blind evaluators - consistency of headings hierarchy, are nav menus
organised as lists, are labels programmatically linked to inputs, are
fieldset/legend or aria group constructs used on non-expressive radio or
checkbox inputs, are data tables correctly marked up, etc. In other
words, there will often be aspects of content where blind testers can
demonstrate failures. It is true that the aspect of 1.3.1 where visual
appearance is checked against semantic markup is not doable, or only to
some extent. For example, in browse mode, the tester may come across
bits of text that could potentially be headings but are not marked up as
such (and then the tester will not know whether they visually stand out
as headings).So I agree that a blind tester alone cannot safely
demonstrate conformance to 1.3.1 but often, he/she can demonstrate
non-conformance.
Detlev

Am 18.09.2019 um 14:12 schrieb Patrick H. Lauke:
> On 18/09/2019 12:30, Steve Green wrote:
>> I don't see how a blind person can test several others, such as 1.3.1
>> Info and Relationships. That SC requires you to assess the visual
>> presentation and check if the information and relationships are
>> conveyed programmatically. How can you do that if you can't see it?
>>
>> There are others like 1.3.2 Meaningful Sequence. If you can only
>> perceive the programmatic reading sequence and not the visual
>> sequence, how can you tell if it is correct?
>>
>> And 1.4.1 Use of Colour. How can you tell if colour alone is being
>> used? It might be possible by analysing the styles, but that won't
>> work if the colour is in images.
>>
>> I won't labour the point, but if you can't perceive something, how
>> can you know if there is anything to perceive?
>
> Oh, absolutely. Sorry, wasn't trying to make an exhaustive list :)
>
> P

--
Detlev Fischer
Testkreis
Werderstr. 34, 20144 Hamburg

Mobil +49 (0)157 57 57 57 45

http://www.testkreis.de
Beratung, Tests und Schulungen für barrierefreie Websites

From: Mark Magennis
Date: Wed, Sep 18 2019 6:37AM
Subject: Re: [EXTERNAL] 1.4.3 & 2.4.7: evaluation while blind
← Previous message | Next message →

Birkir,

I would love if you wrote an article on it. It would be very valuable to a lot of visually impaired developers and QA engineers. Plus, your credentials are perfect for the task and you're good at explaining things.

Mark

Mark Magennis
Skillsoft | mobile: +353 87 60 60 162
Accessibility Specialist

From: Steve Green
Date: Wed, Sep 18 2019 6:53AM
Subject: Re: 1.4.3 & 2.4.7: evaluation while blind
← Previous message | Next message →

I'm afraid I've got to disagree with most of that, Birkir,

You are misunderstanding what 1.3.1 says. It does not say anything about what the heading structure should be. It says that if something visually looks like a heading, it should be marked up as a heading. Visual relationships and information must be conveyed programmatically. The starting point for this is what the content looks like, and you must be able to determine the author's intent. If you can't see it, you can't do that in many cases. I agree that if you can perceive a data table, then you can tell if it is marked up correctly. But if tabular content is marked-up using div or span elements, you may not be able to tell that it should be a table.

1.3.2 is not about a logical content order. It is about identifying the correct content order, which may not be the same. The programmatic content order may well seem logical, yet be incorrect.

I don't want to appear to be the bad guy picking on blind people, but whenever this topic comes up I find that people ignore all the stuff they can't do. It may well be that this is because they simply cannot perceive all those things, so the task of testing appears simpler than it actually is. Depending on what you are testing, it can be extremely difficult even if you are fully able.

Steve

From: Michael Ausbun
Date: Wed, Sep 18 2019 9:17AM
Subject: Re: 1.4.3 & 2.4.7: evaluation while blind
← Previous message | Next message →

I appreciate this; this is how I've always interpreted/done it. Thanks for the concrete suggestions. I have always had the philosophy: "Don't tell me I cannot, let us explore how I can." There are typically methods for doing most things, if I do not have the answer I turn to those who may. Other blind folks with more experience usually have thought of things I haven't.
Thanks again!

From: Patrick H. Lauke
Date: Wed, Sep 18 2019 9:59AM
Subject: Re: 1.4.3 & 2.4.7: evaluation while blind
← Previous message | Next message →

On 18/09/2019 13:31, Detlev Fischer wrote:
> o I agree that a blind tester alone cannot safely
> demonstrate conformance to 1.3.1 but often, he/she can demonstrate
> non-conformance.

Good point, well made.

--
Patrick H. Lauke

www.splintered.co.uk | https://github.com/patrickhlauke
http://flickr.com/photos/redux/ | http://redux.deviantart.com
twitter: @patrick_h_lauke | skype: patrick_h_lauke

From: Birkir R. Gunnarsson
Date: Wed, Sep 18 2019 11:23AM
Subject: Re: 1.4.3 & 2.4.7: evaluation while blind
← Previous message | No next message

Steve

I see your point, and I partially agree, butnot entirely.

You an tell from the content what should potentially be a heading, a
line of text that labels or introduces a section of content.
Then you can check for the presence of either HTML markup such as
<strong> or the presence of a CSS class that includes things like bold
or centered.
If you find either, you already know that the text has visual emphasis.
If you find none, you can still note that this should be marked up as
a heading (if you check with a pair of eyes and notice that there is
no visual emphass at all, you can downgrade your recommendation to a
usability issue, but it's still valid, it's just a matter of
severity).

Yes, you should always check your findings with a sighted person
before you submit

You can identify content that looks like it should be in a table by its nature.
Tabular content is content that needs headers to be understood.

Then you look at the code and see either a presentational/layout table
(table markup with caption or headers) or a series of divs and spans,
the class names typically indicate tabular or a grid-like layout.
Same with lists, if the content appears to be a list of things, it
either has a list markup or it doesn't. If it doesn't you check the
HTML code and class names to see if you can guestimate the author
intent or visual fffect (in most cases, you can).
If it is neither marked up as a list nor preseted aa lit but its
nature is such that it benefits from being presented a list you should
still file an issue, you just have to downgrade it to a usability
recommendation.

I have not come across many places where visual formatting
significantly changes the meaning or nature of content that appears to
be logical to begin ith, and I have audited thousands of webpages.
The most common situation is content that should be hidden for all,
but is only visually hidden, you can figure that out from inspecting
the code.
Typically content that is drastically altered using visual layout
appears to be illogical when inspected with a screen reader.
I'm not sayng there are no such pages, only that they are rare.

Sure,I readily admit to the limitations of being a blind tester (or a
blind person for that matter), I have not applied to pilot school,hair
stylist school or a concealed weapons certificate class, but I think
you somewhat under-estimate the capabilities of a blind tester,
especially if you allow for some extra time to check the underlying
markup.
Your point is entirely legit and I think it's a good discussion to
have (there's no need to get unnecessarily PC about it), so no worries
on that front (if you had any).
I'll see if I can get around to writing this article over the weekend.
It's about time.


On 9/18/19, Patrick H. Lauke < = EMAIL ADDRESS REMOVED = > wrote:
> On 18/09/2019 13:31, Detlev Fischer wrote:
>> o I agree that a blind tester alone cannot safely
>> demonstrate conformance to 1.3.1 but often, he/she can demonstrate
>> non-conformance.
>
> Good point, well made.
>
> --
> Patrick H. Lauke
>
> www.splintered.co.uk | https://github.com/patrickhlauke
> http://flickr.com/photos/redux/ | http://redux.deviantart.com
> twitter: @patrick_h_lauke | skype: patrick_h_lauke
> > > > >


--
Work hard. Have fun. Make history.