Frustrated with NVDA: interferes in the behavior of the page.

classic Classic list List threaded Threaded
17 messages Options
Reply | Threaded
Open this post in threaded view
|

Frustrated with NVDA: interferes in the behavior of the page.

Hedi Carlos Minin
Hello everybody.
I'm a developer and I am studying the development of a tool to deliver tests the visually impaired. However I'm having a lot of frustration with screen readers.
One of these problems is related to javascript and NVDA.
I have the following scenario: when you press the "Enter" key (keydown) on an element, the focus is moved to another element.
But when I'm with open NVDA focus does not move. He somehow interferes in the behavior of the page.
Does anyone know tell me why this happens?
Thank U

------------------------------------------------------------------------------
Dive into the World of Parallel Programming The Go Parallel Website, sponsored
by Intel and developed in partnership with Slashdot Media, is your hub for all
things parallel software development, from weekly thought leadership blogs to
news, videos, case studies, tutorials and more. Take a look and join the
conversation now. http://goparallel.sourceforge.net/
_______________________________________________
Nvda-devel mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/nvda-devel
Reply | Threaded
Open this post in threaded view
|

Re: Frustrated with NVDA: interferes in the behavior of the page.

Aaron Cannon
What type of element is the user pressing enter on? Are you sure that the keydown event is actually being fired when NVDA is on? If pressing enter on a link or button, this will usually generate a click event when the screen reader is active, and not a keypress.

--
This message was sent from a mobile device


> On Mar 29, 2015, at 17:25, Hedi Carlos Minin <[hidden email]> wrote:
>
> Hello everybody.
> I'm a developer and I am studying the development of a tool to deliver tests the visually impaired. However I'm having a lot of frustration with screen readers.
> One of these problems is related to javascript and NVDA.
> I have the following scenario: when you press the "Enter" key (keydown) on an element, the focus is moved to another element.
> But when I'm with open NVDA focus does not move. He somehow interferes in the behavior of the page.
> Does anyone know tell me why this happens?
> Thank U
> ------------------------------------------------------------------------------
> Dive into the World of Parallel Programming The Go Parallel Website, sponsored
> by Intel and developed in partnership with Slashdot Media, is your hub for all
> things parallel software development, from weekly thought leadership blogs to
> news, videos, case studies, tutorials and more. Take a look and join the
> conversation now. http://goparallel.sourceforge.net/
> _______________________________________________
> Nvda-devel mailing list
> [hidden email]
> https://lists.sourceforge.net/lists/listinfo/nvda-devel

------------------------------------------------------------------------------
Dive into the World of Parallel Programming The Go Parallel Website, sponsored
by Intel and developed in partnership with Slashdot Media, is your hub for all
things parallel software development, from weekly thought leadership blogs to
news, videos, case studies, tutorials and more. Take a look and join the
conversation now. http://goparallel.sourceforge.net/
_______________________________________________
Nvda-devel mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/nvda-devel
Reply | Threaded
Open this post in threaded view
|

Re: Frustrated with NVDA: interferes in the behavior of the page.

John Isige
In reply to this post by Hedi Carlos Minin
It's because enter is used to go into forms mode, i.e. the mode you use
to fill in edit boxes and the like. Somebody else here can tell you
about doing things to your page to fix this, but for right now, if you
hit NVDA-F2, that will pass the next keystroke through, and enter will
do what you set it up to do. If I recall correctly, if you set at least
one of the commercial screen readers up to not enter forms mode
automatically, enter also goes into it, so you'll have the same problem
there as well.

On 3/29/2015 17:25, Hedi Carlos Minin wrote:

> Hello everybody.
> I'm a developer and I am studying the development of a tool to deliver
> tests the visually impaired. However I'm having a lot of frustration
> with screen readers.
> One of these problems is related to javascript and NVDA.
> I have the following scenario: when you press the "Enter" key
> (keydown) on an element, the focus is moved to another element.
> But when I'm with open NVDA focus does not move. He somehow interferes
> in the behavior of the page.
> Does anyone know tell me why this happens?
> Thank U
>
>
> ------------------------------------------------------------------------------
> Dive into the World of Parallel Programming The Go Parallel Website, sponsored
> by Intel and developed in partnership with Slashdot Media, is your hub for all
> things parallel software development, from weekly thought leadership blogs to
> news, videos, case studies, tutorials and more. Take a look and join the
> conversation now. http://goparallel.sourceforge.net/
>
>
> _______________________________________________
> Nvda-devel mailing list
> [hidden email]
> https://lists.sourceforge.net/lists/listinfo/nvda-devel


------------------------------------------------------------------------------
Dive into the World of Parallel Programming The Go Parallel Website, sponsored
by Intel and developed in partnership with Slashdot Media, is your hub for all
things parallel software development, from weekly thought leadership blogs to
news, videos, case studies, tutorials and more. Take a look and join the
conversation now. http://goparallel.sourceforge.net/
_______________________________________________
Nvda-devel mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/nvda-devel
Reply | Threaded
Open this post in threaded view
|

Re: Frustrated with NVDA: interferes in the behavior of the page.

Hedi Carlos Minin
Thank you guys. That's right that's going on.
Have "a" elements with the role attribute = "presentation" (do not want NVDA announce the element, only its contents), with the keydown event and click (the application is not only for the blind).

When NVDA is on the keydown event is not fired. It happens to the key space as well. But I believe the assistive technology should not interfere with the application behavior.

Unfortunately this is not the only problem I'm having. When, for example, I change the content of an element and move the focus to it, NVDA reads the old content. Even trying to use the tags "aria-live" does not work.

I must confess that you get to be discouraged, because I need to develop an application for the blind (and make the best possible), but there are many problems: different browsers with different tlas readers have different behaviors. For example NVDA does not announce "clickable" when I have an "a" element with the "click" event in Chrome, but Firefox does.

2015-03-29 20:50 GMT-03:00 John Schucker <[hidden email]>:
It's because enter is used to go into forms mode, i.e. the mode you use
to fill in edit boxes and the like. Somebody else here can tell you
about doing things to your page to fix this, but for right now, if you
hit NVDA-F2, that will pass the next keystroke through, and enter will
do what you set it up to do. If I recall correctly, if you set at least
one of the commercial screen readers up to not enter forms mode
automatically, enter also goes into it, so you'll have the same problem
there as well.

On 3/29/2015 17:25, Hedi Carlos Minin wrote:
> Hello everybody.
> I'm a developer and I am studying the development of a tool to deliver
> tests the visually impaired. However I'm having a lot of frustration
> with screen readers.
> One of these problems is related to javascript and NVDA.
> I have the following scenario: when you press the "Enter" key
> (keydown) on an element, the focus is moved to another element.
> But when I'm with open NVDA focus does not move. He somehow interferes
> in the behavior of the page.
> Does anyone know tell me why this happens?
> Thank U
>
>
> ------------------------------------------------------------------------------
> Dive into the World of Parallel Programming The Go Parallel Website, sponsored
> by Intel and developed in partnership with Slashdot Media, is your hub for all
> things parallel software development, from weekly thought leadership blogs to
> news, videos, case studies, tutorials and more. Take a look and join the
> conversation now. http://goparallel.sourceforge.net/
>
>
> _______________________________________________
> Nvda-devel mailing list
> [hidden email]
> https://lists.sourceforge.net/lists/listinfo/nvda-devel


------------------------------------------------------------------------------
Dive into the World of Parallel Programming The Go Parallel Website, sponsored
by Intel and developed in partnership with Slashdot Media, is your hub for all
things parallel software development, from weekly thought leadership blogs to
news, videos, case studies, tutorials and more. Take a look and join the
conversation now. http://goparallel.sourceforge.net/
_______________________________________________
Nvda-devel mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/nvda-devel


------------------------------------------------------------------------------
Dive into the World of Parallel Programming The Go Parallel Website, sponsored
by Intel and developed in partnership with Slashdot Media, is your hub for all
things parallel software development, from weekly thought leadership blogs to
news, videos, case studies, tutorials and more. Take a look and join the
conversation now. http://goparallel.sourceforge.net/
_______________________________________________
Nvda-devel mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/nvda-devel
Reply | Threaded
Open this post in threaded view
|

Re: Frustrated with NVDA: interferes in the behavior of the page.

Hedi Carlos Minin
As this is an application for test delivery, where the candidate is being evaluated, I need to have a minimum of speech (just enough). I do not need to announce a list nor its entirety, because the list of options will always have 5 elements (A, B, C, D and E). The announcement of the elements can cause confusion, for example, if I have the list:

A: 2.
B: 4.
C: 7.
D: 9
E: 11

and NVDA reads:

A: 2 1 of 5
B: 4 2 of 5
C: 7 3 of 5
D: 9 4 of 5
E: 11 5 of 5

The memory is very short and this additional talk can be distracting. This is a very specific application and I need to have total control over what is or is not spoken. But I'm realizing that this is not an easy task.

------------------------------------------------------------------------------
Dive into the World of Parallel Programming The Go Parallel Website, sponsored
by Intel and developed in partnership with Slashdot Media, is your hub for all
things parallel software development, from weekly thought leadership blogs to
news, videos, case studies, tutorials and more. Take a look and join the
conversation now. http://goparallel.sourceforge.net/
_______________________________________________
Nvda-devel mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/nvda-devel
Reply | Threaded
Open this post in threaded view
|

Re: Frustrated with NVDA: interferes in the behavior of the page.

James Teh
In reply to this post by Hedi Carlos Minin
On 30/03/2015 10:37 AM, Hedi Carlos Minin wrote:
> When NVDA is on the keydown event is not fired. It happens to the key
> space as well. But I believe the assistive technology should not
> interfere with the application behavior.
That's a great principle in theory and we try to honour it as much as
possible. Unfortunately, not every interactive element on the web is
focusable and/or is keyboard accessible, so we have to provide a
standard way to activate elements in browse mode. Whether we like it or
not (and I don't :) ), the one reliable way on the web is clicking. We
have no way of knowing whether an element is keyboard accessible or not,
so we can't make exceptions for those cases.

> Unfortunately this is not the only problem I'm having. When, for
> example, I change the content of an element and move the focus to it,
> NVDA reads the old content.
Please provide a test case.

> For example NVDA does not announce "clickable" when
> I have an "a" element with the "click" event in Chrome, but Firefox does.
I recommend you avoid relying on "clickable" as a way to indicate that a
user can activate something. Essentially, the fact that we say
"clickable" at all is a major hack to allow users to activate elements
that aren't properly accessible in the first place (e.g. a clickable
span without an ARIA role). In an ideal world, we'd never need to say
"clickable" at all.

That said, the reason for the different behaviour here is that Chrome
doesn't expose the information we need to report this.

Jamie

--
James Teh
Executive Director, NV Access Limited
Ph +61 7 3149 3306
www.nvaccess.org
Facebook: http://www.facebook.com/NVAccess
Twitter: @NVAccess
SIP: [hidden email]

------------------------------------------------------------------------------
Dive into the World of Parallel Programming The Go Parallel Website, sponsored
by Intel and developed in partnership with Slashdot Media, is your hub for all
things parallel software development, from weekly thought leadership blogs to
news, videos, case studies, tutorials and more. Take a look and join the
conversation now. http://goparallel.sourceforge.net/
_______________________________________________
Nvda-devel mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/nvda-devel
Reply | Threaded
Open this post in threaded view
|

Re: Frustrated with NVDA: interferes in the behavior of the page.

Hedi Carlos Minin
Thanks Jamie.

In fact I do not want him to say "clickable", because users will be informed that you must press enter to select one of the options (remember that we want to reduce the information load).

For simplicity, the page has few elements and navigation can be performed using only "Tab" and "Shift + Tab". We assign the attribute "tabindex" to all elements to be accessible to the keyboard.

Here is the link of the page (in Portuguese, sorry):


Hedi

2015-03-29 23:51 GMT-03:00 James Teh <[hidden email]>:
On 30/03/2015 10:37 AM, Hedi Carlos Minin wrote:
> When NVDA is on the keydown event is not fired. It happens to the key
> space as well. But I believe the assistive technology should not
> interfere with the application behavior.
That's a great principle in theory and we try to honour it as much as
possible. Unfortunately, not every interactive element on the web is
focusable and/or is keyboard accessible, so we have to provide a
standard way to activate elements in browse mode. Whether we like it or
not (and I don't :) ), the one reliable way on the web is clicking. We
have no way of knowing whether an element is keyboard accessible or not,
so we can't make exceptions for those cases.

> Unfortunately this is not the only problem I'm having. When, for
> example, I change the content of an element and move the focus to it,
> NVDA reads the old content.
Please provide a test case.

> For example NVDA does not announce "clickable" when
> I have an "a" element with the "click" event in Chrome, but Firefox does.
I recommend you avoid relying on "clickable" as a way to indicate that a
user can activate something. Essentially, the fact that we say
"clickable" at all is a major hack to allow users to activate elements
that aren't properly accessible in the first place (e.g. a clickable
span without an ARIA role). In an ideal world, we'd never need to say
"clickable" at all.

That said, the reason for the different behaviour here is that Chrome
doesn't expose the information we need to report this.

Jamie

--
James Teh
Executive Director, NV Access Limited
Ph <a href="tel:%2B61%207%203149%203306" value="+61731493306">+61 7 3149 3306
www.nvaccess.org
Facebook: http://www.facebook.com/NVAccess
Twitter: @NVAccess
SIP: [hidden email]

------------------------------------------------------------------------------
Dive into the World of Parallel Programming The Go Parallel Website, sponsored
by Intel and developed in partnership with Slashdot Media, is your hub for all
things parallel software development, from weekly thought leadership blogs to
news, videos, case studies, tutorials and more. Take a look and join the
conversation now. http://goparallel.sourceforge.net/
_______________________________________________
Nvda-devel mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/nvda-devel


------------------------------------------------------------------------------
Dive into the World of Parallel Programming The Go Parallel Website, sponsored
by Intel and developed in partnership with Slashdot Media, is your hub for all
things parallel software development, from weekly thought leadership blogs to
news, videos, case studies, tutorials and more. Take a look and join the
conversation now. http://goparallel.sourceforge.net/
_______________________________________________
Nvda-devel mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/nvda-devel
Reply | Threaded
Open this post in threaded view
|

Re: Frustrated with NVDA: interferes in the behavior of the page.

Hedi Carlos Minin
You may wonder, why not utilized checkbox? because I had problems with it also in Firefox. And it does not increase with the zoom on some browsers.

I also tried using the "application role" in the page body, but the problems only increased.

In application mode, NVDA reads the contents of the aria-label attribute (or aria-labelledby), but if I use the text for navigation shortcuts (talk word for word for example) NVDA reads the element content (and not the content of the aria-label attribute). I do not know if this is normal, but if the contents are different there is no consistency.

2015-03-30 0:38 GMT-03:00 Hedi Carlos Minin <[hidden email]>:
Thanks Jamie.

In fact I do not want him to say "clickable", because users will be informed that you must press enter to select one of the options (remember that we want to reduce the information load).

For simplicity, the page has few elements and navigation can be performed using only "Tab" and "Shift + Tab". We assign the attribute "tabindex" to all elements to be accessible to the keyboard.

Here is the link of the page (in Portuguese, sorry):


Hedi

2015-03-29 23:51 GMT-03:00 James Teh <[hidden email]>:

On 30/03/2015 10:37 AM, Hedi Carlos Minin wrote:
> When NVDA is on the keydown event is not fired. It happens to the key
> space as well. But I believe the assistive technology should not
> interfere with the application behavior.
That's a great principle in theory and we try to honour it as much as
possible. Unfortunately, not every interactive element on the web is
focusable and/or is keyboard accessible, so we have to provide a
standard way to activate elements in browse mode. Whether we like it or
not (and I don't :) ), the one reliable way on the web is clicking. We
have no way of knowing whether an element is keyboard accessible or not,
so we can't make exceptions for those cases.

> Unfortunately this is not the only problem I'm having. When, for
> example, I change the content of an element and move the focus to it,
> NVDA reads the old content.
Please provide a test case.

> For example NVDA does not announce "clickable" when
> I have an "a" element with the "click" event in Chrome, but Firefox does.
I recommend you avoid relying on "clickable" as a way to indicate that a
user can activate something. Essentially, the fact that we say
"clickable" at all is a major hack to allow users to activate elements
that aren't properly accessible in the first place (e.g. a clickable
span without an ARIA role). In an ideal world, we'd never need to say
"clickable" at all.

That said, the reason for the different behaviour here is that Chrome
doesn't expose the information we need to report this.

Jamie

--
James Teh
Executive Director, NV Access Limited
Ph <a href="tel:%2B61%207%203149%203306" value="+61731493306" target="_blank">+61 7 3149 3306
www.nvaccess.org
Facebook: http://www.facebook.com/NVAccess
Twitter: @NVAccess
SIP: [hidden email]

------------------------------------------------------------------------------
Dive into the World of Parallel Programming The Go Parallel Website, sponsored
by Intel and developed in partnership with Slashdot Media, is your hub for all
things parallel software development, from weekly thought leadership blogs to
news, videos, case studies, tutorials and more. Take a look and join the
conversation now. http://goparallel.sourceforge.net/
_______________________________________________
Nvda-devel mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/nvda-devel



------------------------------------------------------------------------------
Dive into the World of Parallel Programming The Go Parallel Website, sponsored
by Intel and developed in partnership with Slashdot Media, is your hub for all
things parallel software development, from weekly thought leadership blogs to
news, videos, case studies, tutorials and more. Take a look and join the
conversation now. http://goparallel.sourceforge.net/
_______________________________________________
Nvda-devel mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/nvda-devel
Reply | Threaded
Open this post in threaded view
|

Re: Frustrated with NVDA: interferes in the behavior of the page.

Rui Fontes
In reply to this post by Hedi Carlos Minin
If it is for a specific application, why not create a profil where the announcement of the position is disabled?
 
Rui
 
 
Data: segunda-feira, 30 de Março de 2015 01:46
Assunto: Re: [Nvda-devel] Frustrated with NVDA: interferes in the behavior of the page.
 
As this is an application for test delivery, where the candidate is being evaluated, I need to have a minimum of speech (just enough). I do not need to announce a list nor its entirety, because the list of options will always have 5 elements (A, B, C, D and E). The announcement of the elements can cause confusion, for example, if I have the list:
 
A: 2.
B: 4.
C: 7.
D: 9
E: 11
 
and NVDA reads:
 
A: 2 1 of 5
B: 4 2 of 5
C: 7 3 of 5
D: 9 4 of 5
E: 11 5 of 5
 
The memory is very short and this additional talk can be distracting. This is a very specific application and I need to have total control over what is or is not spoken. But I'm realizing that this is not an easy task.


------------------------------------------------------------------------------
Dive into the World of Parallel Programming The Go Parallel Website, sponsored
by Intel and developed in partnership with Slashdot Media, is your hub for all
things parallel software development, from weekly thought leadership blogs to
news, videos, case studies, tutorials and more. Take a look and join the
conversation now. http://goparallel.sourceforge.net/


_______________________________________________
Nvda-devel mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/nvda-devel

------------------------------------------------------------------------------
Dive into the World of Parallel Programming The Go Parallel Website, sponsored
by Intel and developed in partnership with Slashdot Media, is your hub for all
things parallel software development, from weekly thought leadership blogs to
news, videos, case studies, tutorials and more. Take a look and join the
conversation now. http://goparallel.sourceforge.net/
_______________________________________________
Nvda-devel mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/nvda-devel
Reply | Threaded
Open this post in threaded view
|

Re: Frustrated with NVDA: interferes in the behavior of the page.

Aaron Cannon
The big problem, in my mind, with what you're doing is you are trying
to make decisions about what a user will and will not expect to hear.
Screen reader users know that you are supposed to press enter to
click.  Screen reader users are used to hearing lists, and should not
be surprised when they are told that a list of answers is in fact one.

In my experience, one of the hardest things for persons new to web
accessibility is knowing when a screen reader is doing something
expected or not.  Things that can seem like a bug are often not, and
vice versa.

Also, if you are working with Chrome, you should know that you are
working with the browser, which, I believe, is the least well
supported by NVDA, or pretty much any windows screen reader.  If you
have a choice, I'd recommend Firefox and/or IE, with my preferred
browser being FF.

If I were you, unless you have some special requirements that you
haven't shared, I'd just trust your users to figure it out, and just
mark up your page with standard HTML controls.  Make your answers
radio buttons or checkboxes (if multiple answers are permitted).
Trust your user to know that they can press enter to click a link or
button.  Otherwise, it seems to me that you may be attempting to
reinvent your own screen reader on top of an existing one.

If you really do need complete control over exactly what is spoken,
then I fear your only option may be to create a self-voicing
application, something which is possible but very fragile as a web
application.  Most self-voicing apps are native applications.

Good luck.

Aaron

On 3/30/15, Rui Fontes <[hidden email]> wrote:

> If it is for a specific application, why not create a profil where the
> announcement of the position is disabled?
>
> Rui
>
>
> De: Hedi Carlos Minin
> Data: segunda-feira, 30 de Março de 2015 01:46
> Para: NVDA screen reader development
> Assunto: Re: [Nvda-devel] Frustrated with NVDA: interferes in the behavior
> of the page.
>
> As this is an application for test delivery, where the candidate is being
> evaluated, I need to have a minimum of speech (just enough). I do not need
> to announce a list nor its entirety, because the list of options will always
> have 5 elements (A, B, C, D and E). The announcement of the elements can
> cause confusion, for example, if I have the list:
>
> A: 2.
> B: 4.
> C: 7.
> D: 9
> E: 11
>
> and NVDA reads:
>
> A: 2 1 of 5
> B: 4 2 of 5
>
> C: 7 3 of 5
>
> D: 9 4 of 5
>
> E: 11 5 of 5
>
> The memory is very short and this additional talk can be distracting. This
> is a very specific application and I need to have total control over what is
> or is not spoken. But I'm realizing that this is not an easy task.
>
>
>
> --------------------------------------------------------------------------------
> ------------------------------------------------------------------------------
> Dive into the World of Parallel Programming The Go Parallel Website,
> sponsored
> by Intel and developed in partnership with Slashdot Media, is your hub for
> all
> things parallel software development, from weekly thought leadership blogs
> to
> news, videos, case studies, tutorials and more. Take a look and join the
> conversation now. http://goparallel.sourceforge.net/
>
>
> --------------------------------------------------------------------------------
> _______________________________________________
> Nvda-devel mailing list
> [hidden email]
> https://lists.sourceforge.net/lists/listinfo/nvda-devel
>

------------------------------------------------------------------------------
Dive into the World of Parallel Programming The Go Parallel Website, sponsored
by Intel and developed in partnership with Slashdot Media, is your hub for all
things parallel software development, from weekly thought leadership blogs to
news, videos, case studies, tutorials and more. Take a look and join the
conversation now. http://goparallel.sourceforge.net/
_______________________________________________
Nvda-devel mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/nvda-devel
Reply | Threaded
Open this post in threaded view
|

Re: Frustrated with NVDA: interferes in the behavior of the page.

Hedi Carlos Minin
Aaron, I agree with you on some points.
We want to use screen readers because users are already accustomed to using and we should consider it.
Navigation should be simple: a few elements and able to go through each element of the application with Tab and Shift + Tab.

A question has a clear format: a few paragraphs containing the description and five response options. There are 90 questions, then announce "paragraph" after each paragraph and "list item" after each response option becomes redundant, tiring and can cause confusion on some question. Only are these two elements I am avoiding the speech of the role.
Of course, a button must be advertised as a button, so that users know you can perform an action.

The application is also for people with low vision. The problem I encountered when using standard radio controls is that it does not increase with the zoom on some browsers (like Firefox).

I had considered a desktop application, but display mathml, svg, and have control over zoom, constrast, is not a trivial task. For that HTML 5 is an interesting option.

I also considered one self voice application. In JavaFX I have a Web browser, and have control over various aspects. My problem is that not only hear the questions, you need to write well. I would have to "emulate" a screen reader, which is not an easy task.

Well I think I should choose a browser (such as Firefox) and design the application to be compatible with it.

Hedi

2015-03-30 10:16 GMT-03:00 Aaron Cannon <[hidden email]>:
The big problem, in my mind, with what you're doing is you are trying
to make decisions about what a user will and will not expect to hear.
Screen reader users know that you are supposed to press enter to
click.  Screen reader users are used to hearing lists, and should not
be surprised when they are told that a list of answers is in fact one.

In my experience, one of the hardest things for persons new to web
accessibility is knowing when a screen reader is doing something
expected or not.  Things that can seem like a bug are often not, and
vice versa.

Also, if you are working with Chrome, you should know that you are
working with the browser, which, I believe, is the least well
supported by NVDA, or pretty much any windows screen reader.  If you
have a choice, I'd recommend Firefox and/or IE, with my preferred
browser being FF.

If I were you, unless you have some special requirements that you
haven't shared, I'd just trust your users to figure it out, and just
mark up your page with standard HTML controls.  Make your answers
radio buttons or checkboxes (if multiple answers are permitted).
Trust your user to know that they can press enter to click a link or
button.  Otherwise, it seems to me that you may be attempting to
reinvent your own screen reader on top of an existing one.

If you really do need complete control over exactly what is spoken,
then I fear your only option may be to create a self-voicing
application, something which is possible but very fragile as a web
application.  Most self-voicing apps are native applications.

Good luck.

Aaron

On 3/30/15, Rui Fontes <[hidden email]> wrote:
> If it is for a specific application, why not create a profil where the
> announcement of the position is disabled?
>
> Rui
>
>
> De: Hedi Carlos Minin
> Data: segunda-feira, 30 de Março de 2015 01:46
> Para: NVDA screen reader development
> Assunto: Re: [Nvda-devel] Frustrated with NVDA: interferes in the behavior
> of the page.
>
> As this is an application for test delivery, where the candidate is being
> evaluated, I need to have a minimum of speech (just enough). I do not need
> to announce a list nor its entirety, because the list of options will always
> have 5 elements (A, B, C, D and E). The announcement of the elements can
> cause confusion, for example, if I have the list:
>
> A: 2.
> B: 4.
> C: 7.
> D: 9
> E: 11
>
> and NVDA reads:
>
> A: 2 1 of 5
> B: 4 2 of 5
>
> C: 7 3 of 5
>
> D: 9 4 of 5
>
> E: 11 5 of 5
>
> The memory is very short and this additional talk can be distracting. This
> is a very specific application and I need to have total control over what is
> or is not spoken. But I'm realizing that this is not an easy task.
>
>
>
> --------------------------------------------------------------------------------
> ------------------------------------------------------------------------------
> Dive into the World of Parallel Programming The Go Parallel Website,
> sponsored
> by Intel and developed in partnership with Slashdot Media, is your hub for
> all
> things parallel software development, from weekly thought leadership blogs
> to
> news, videos, case studies, tutorials and more. Take a look and join the
> conversation now. http://goparallel.sourceforge.net/
>
>
> --------------------------------------------------------------------------------
> _______________________________________________
> Nvda-devel mailing list
> [hidden email]
> https://lists.sourceforge.net/lists/listinfo/nvda-devel
>

------------------------------------------------------------------------------
Dive into the World of Parallel Programming The Go Parallel Website, sponsored
by Intel and developed in partnership with Slashdot Media, is your hub for all
things parallel software development, from weekly thought leadership blogs to
news, videos, case studies, tutorials and more. Take a look and join the
conversation now. http://goparallel.sourceforge.net/
_______________________________________________
Nvda-devel mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/nvda-devel


------------------------------------------------------------------------------
Dive into the World of Parallel Programming The Go Parallel Website, sponsored
by Intel and developed in partnership with Slashdot Media, is your hub for all
things parallel software development, from weekly thought leadership blogs to
news, videos, case studies, tutorials and more. Take a look and join the
conversation now. http://goparallel.sourceforge.net/
_______________________________________________
Nvda-devel mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/nvda-devel
Reply | Threaded
Open this post in threaded view
|

Re: Frustrated with NVDA: interferes in the behavior of the page.

mk360
Ok, but NVDA users know what will be anounced or not when reads one webpage. We know that paragraph will be anounced and if not that can be confusing for a user that uses usually screen readers that can appear to be redundant to you. Also the different anouncements can be enabled or dissabled, so the user ob the screen reader decides if it needs to be anounced or not. for example, the user can disable onclic anouncement on the formating options or similar (formateado de documentos in my NVDA).

Regards,
mk.

On Mon, Mar 30, 2015 at 11:13 AM, Hedi Carlos Minin <[hidden email]> wrote:
Aaron, I agree with you on some points.
We want to use screen readers because users are already accustomed to using and we should consider it.
Navigation should be simple: a few elements and able to go through each element of the application with Tab and Shift + Tab.

A question has a clear format: a few paragraphs containing the description and five response options. There are 90 questions, then announce "paragraph" after each paragraph and "list item" after each response option becomes redundant, tiring and can cause confusion on some question. Only are these two elements I am avoiding the speech of the role.
Of course, a button must be advertised as a button, so that users know you can perform an action.

The application is also for people with low vision. The problem I encountered when using standard radio controls is that it does not increase with the zoom on some browsers (like Firefox).

I had considered a desktop application, but display mathml, svg, and have control over zoom, constrast, is not a trivial task. For that HTML 5 is an interesting option.

I also considered one self voice application. In JavaFX I have a Web browser, and have control over various aspects. My problem is that not only hear the questions, you need to write well. I would have to "emulate" a screen reader, which is not an easy task.

Well I think I should choose a browser (such as Firefox) and design the application to be compatible with it.

Hedi

2015-03-30 10:16 GMT-03:00 Aaron Cannon <[hidden email]>:

The big problem, in my mind, with what you're doing is you are trying
to make decisions about what a user will and will not expect to hear.
Screen reader users know that you are supposed to press enter to
click.  Screen reader users are used to hearing lists, and should not
be surprised when they are told that a list of answers is in fact one.

In my experience, one of the hardest things for persons new to web
accessibility is knowing when a screen reader is doing something
expected or not.  Things that can seem like a bug are often not, and
vice versa.

Also, if you are working with Chrome, you should know that you are
working with the browser, which, I believe, is the least well
supported by NVDA, or pretty much any windows screen reader.  If you
have a choice, I'd recommend Firefox and/or IE, with my preferred
browser being FF.

If I were you, unless you have some special requirements that you
haven't shared, I'd just trust your users to figure it out, and just
mark up your page with standard HTML controls.  Make your answers
radio buttons or checkboxes (if multiple answers are permitted).
Trust your user to know that they can press enter to click a link or
button.  Otherwise, it seems to me that you may be attempting to
reinvent your own screen reader on top of an existing one.

If you really do need complete control over exactly what is spoken,
then I fear your only option may be to create a self-voicing
application, something which is possible but very fragile as a web
application.  Most self-voicing apps are native applications.

Good luck.

Aaron

On 3/30/15, Rui Fontes <[hidden email]> wrote:
> If it is for a specific application, why not create a profil where the
> announcement of the position is disabled?
>
> Rui
>
>
> De: Hedi Carlos Minin
> Data: segunda-feira, 30 de Março de 2015 01:46
> Para: NVDA screen reader development
> Assunto: Re: [Nvda-devel] Frustrated with NVDA: interferes in the behavior
> of the page.
>
> As this is an application for test delivery, where the candidate is being
> evaluated, I need to have a minimum of speech (just enough). I do not need
> to announce a list nor its entirety, because the list of options will always
> have 5 elements (A, B, C, D and E). The announcement of the elements can
> cause confusion, for example, if I have the list:
>
> A: 2.
> B: 4.
> C: 7.
> D: 9
> E: 11
>
> and NVDA reads:
>
> A: 2 1 of 5
> B: 4 2 of 5
>
> C: 7 3 of 5
>
> D: 9 4 of 5
>
> E: 11 5 of 5
>
> The memory is very short and this additional talk can be distracting. This
> is a very specific application and I need to have total control over what is
> or is not spoken. But I'm realizing that this is not an easy task.
>
>
>
> --------------------------------------------------------------------------------
> ------------------------------------------------------------------------------
> Dive into the World of Parallel Programming The Go Parallel Website,
> sponsored
> by Intel and developed in partnership with Slashdot Media, is your hub for
> all
> things parallel software development, from weekly thought leadership blogs
> to
> news, videos, case studies, tutorials and more. Take a look and join the
> conversation now. http://goparallel.sourceforge.net/
>
>
> --------------------------------------------------------------------------------
> _______________________________________________
> Nvda-devel mailing list
> [hidden email]
> https://lists.sourceforge.net/lists/listinfo/nvda-devel
>

------------------------------------------------------------------------------
Dive into the World of Parallel Programming The Go Parallel Website, sponsored
by Intel and developed in partnership with Slashdot Media, is your hub for all
things parallel software development, from weekly thought leadership blogs to
news, videos, case studies, tutorials and more. Take a look and join the
conversation now. http://goparallel.sourceforge.net/
_______________________________________________
Nvda-devel mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/nvda-devel


------------------------------------------------------------------------------
Dive into the World of Parallel Programming The Go Parallel Website, sponsored
by Intel and developed in partnership with Slashdot Media, is your hub for all
things parallel software development, from weekly thought leadership blogs to
news, videos, case studies, tutorials and more. Take a look and join the
conversation now. http://goparallel.sourceforge.net/
_______________________________________________
Nvda-devel mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/nvda-devel



------------------------------------------------------------------------------
Dive into the World of Parallel Programming The Go Parallel Website, sponsored
by Intel and developed in partnership with Slashdot Media, is your hub for all
things parallel software development, from weekly thought leadership blogs to
news, videos, case studies, tutorials and more. Take a look and join the
conversation now. http://goparallel.sourceforge.net/
_______________________________________________
Nvda-devel mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/nvda-devel
Reply | Threaded
Open this post in threaded view
|

Re: Frustrated with NVDA: interferes in the behavior of the page.

Doug Lee
In reply to this post by Hedi Carlos Minin
------------------------------------------------------------------------------
Dive into the World of Parallel Programming The Go Parallel Website, sponsored
by Intel and developed in partnership with Slashdot Media, is your hub for all
things parallel software development, from weekly thought leadership blogs to
news, videos, case studies, tutorials and more. Take a look and join the
conversation now. http://goparallel.sourceforge.net/
_______________________________________________
Nvda-devel mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/nvda-devel
Reply | Threaded
Open this post in threaded view
|

Re: Frustrated with NVDA: interferes in the behavior of the page.

Hedi Carlos Minin
Thank you guys, I'm glad to respond.
I'm learning a lot.
It seems I'm being boring insisting on something that will not work, sorry.

The objective of this work is to identify requirements for delivery of an examination applied every year in Brazil for middle school students through the computer for the blind.

After all these answers I realize that the best thing to do is create the complication using standard elements as much as possible and not try to change them, leaving the user in control.

I will perform usability tests, then I can test two applications and ask the user which you prefer.

This is my first work specifically for the visually impaired, and I realize that there are many doubts about many details. Every day I learn a lot.

Sorry for my English, is Google translate.


2015-03-30 11:25 GMT-03:00 Doug Lee <[hidden email]>:
Hedi,

James is in Australia and thus not likely to respond again for several
hours, if I had to guess. I would consider him an authority in here on
things like this. That said though...

I have been professionally adapting screen readers to work with
applications for over 15 years, and I thought it might be good for me
to weigh in...

A blind person gets used to speech patterns much as a sighted person
gets used to visual ones. When the pattern changes, in both cases, we
naturally wonder why, and may sometimes even be confused by the
change.

So, for a list, the absence of "1 of 5," etc. might make a blind
person wonder whether this is really a list. Of course in context, it
may quickly become obvious for your situation; but initially, it may
be a sort of confusion that could be avoided.

I don't presume to give you the right way to code this; I simply
suggest that the reactions you get from blind users on this list may
help you see more clearly what really could confuse us, based on what
we have come over many years to expect.

I hope this is helpful, and I am very appreciative that you have come
to a list composed of blind users and developers to discuss how best
to develop something we are meant to use. It could go without saying,
but it won't: We would love to see more of this! :-)

On Mon, Mar 30, 2015 at 11:13:42AM -0300, Hedi Carlos Minin wrote:
   Aaron, I agree with you on some points.
   We want to use screen readers because users are already accustomed to
   using and we should consider it.
   Navigation should be simple: a few elements and able to go through each
   element of the application with Tab and Shift + Tab.
   A question has a clear format: a few paragraphs containing the
   description and five response options. There are 90 questions, then
   announce "paragraph" after each paragraph and "list item" after each
   response option becomes redundant, tiring and can cause confusion on
   some question. Only are these two elements I am avoiding the speech of
   the role.
   Of course, a button must be advertised as a button, so that users know
   you can perform an action.
   The application is also for people with low vision. The problem I
   encountered when using standard radio controls is that it does not
   increase with the zoom on some browsers (like Firefox).
   I had considered a desktop application, but display mathml, svg, and
   have control over zoom, constrast, is not a trivial task. For that HTML
   5 is an interesting option.
   I also considered one self voice application. In JavaFX I have a Web
   browser, and have control over various aspects. My problem is that not
   only hear the questions, you need to write well. I would have to
   "emulate" a screen reader, which is not an easy task.
   Well I think I should choose a browser (such as Firefox) and design the
   application to be compatible with it.
   Hedi

   2015-03-30 10:16 GMT-03:00 Aaron Cannon
   <[1][hidden email]>:

     The big problem, in my mind, with what you're doing is you are
     trying
     to make decisions about what a user will and will not expect to
     hear.
     Screen reader users know that you are supposed to press enter to
     click.  Screen reader users are used to hearing lists, and should
     not
     be surprised when they are told that a list of answers is in fact
     one.
     In my experience, one of the hardest things for persons new to web
     accessibility is knowing when a screen reader is doing something
     expected or not.  Things that can seem like a bug are often not,
     and
     vice versa.
     Also, if you are working with Chrome, you should know that you are
     working with the browser, which, I believe, is the least well
     supported by NVDA, or pretty much any windows screen reader.  If
     you
     have a choice, I'd recommend Firefox and/or IE, with my preferred
     browser being FF.
     If I were you, unless you have some special requirements that you
     haven't shared, I'd just trust your users to figure it out, and just
     mark up your page with standard HTML controls.  Make your answers
     radio buttons or checkboxes (if multiple answers are permitted).
     Trust your user to know that they can press enter to click a link or
     button.  Otherwise, it seems to me that you may be attempting to
     reinvent your own screen reader on top of an existing one.
     If you really do need complete control over exactly what is spoken,
     then I fear your only option may be to create a self-voicing
     application, something which is possible but very fragile as a web
     application.  Most self-voicing apps are native applications.
     Good luck.
     Aaron

   On 3/30/15, Rui Fontes <[2][hidden email]> wrote:
   > If it is for a specific application, why not create a profil where
   the
   > announcement of the position is disabled?
   >
   > Rui
   >
   >
   > De: Hedi Carlos Minin
   > Data: segunda-feira, 30 de Março de 2015 01:46
   > Para: NVDA screen reader development
   > Assunto: Re: [Nvda-devel] Frustrated with NVDA: interferes in the
   behavior
   > of the page.
   >
   > As this is an application for test delivery, where the candidate is
   being
   > evaluated, I need to have a minimum of speech (just enough). I do not
   need
   > to announce a list nor its entirety, because the list of options will
   always
   > have 5 elements (A, B, C, D and E). The announcement of the elements
   can
   > cause confusion, for example, if I have the list:
   >
   > A: 2.
   > B: 4.
   > C: 7.
   > D: 9
   > E: 11
   >
   > and NVDA reads:
   >
   > A: 2 1 of 5
   > B: 4 2 of 5
   >
   > C: 7 3 of 5
   >
   > D: 9 4 of 5
   >
   > E: 11 5 of 5
   >
   > The memory is very short and this additional talk can be distracting.
   This
   > is a very specific application and I need to have total control over
   what is
   > or is not spoken. But I'm realizing that this is not an easy task.
   >
   >
   >

     >
     --------------------------------------------------------------------
     ------------
     >
     --------------------------------------------------------------------
     ----------
     > Dive into the World of Parallel Programming The Go Parallel
     Website,
     > sponsored
     > by Intel and developed in partnership with Slashdot Media, is your
     hub for
     > all
     > things parallel software development, from weekly thought
     leadership blogs
     > to
     > news, videos, case studies, tutorials and more. Take a look and
     join the
     > conversation now. [3]http://goparallel.sourceforge.net/
     >
     >
     >
     --------------------------------------------------------------------
     ------------
     > _______________________________________________
     > Nvda-devel mailing list
     > [4][hidden email]
     > [5]https://lists.sourceforge.net/lists/listinfo/nvda-devel
     >
     --------------------------------------------------------------------
     ----------
     Dive into the World of Parallel Programming The Go Parallel Website,
     sponsored
     by Intel and developed in partnership with Slashdot Media, is your
     hub for all
     things parallel software development, from weekly thought leadership
     blogs to
     news, videos, case studies, tutorials and more. Take a look and join
     the
     conversation now. [6]http://goparallel.sourceforge.net/

   _______________________________________________
   Nvda-devel mailing list
   [7][hidden email]
   [8]https://lists.sourceforge.net/lists/listinfo/nvda-devel

References

   1. mailto:[hidden email]
   2. mailto:[hidden email]
   3. http://goparallel.sourceforge.net/
   4. mailto:[hidden email]
   5. https://lists.sourceforge.net/lists/listinfo/nvda-devel
   6. http://goparallel.sourceforge.net/
   7. mailto:[hidden email]
   8. https://lists.sourceforge.net/lists/listinfo/nvda-devel

------------------------------------------------------------------------------
Dive into the World of Parallel Programming The Go Parallel Website, sponsored
by Intel and developed in partnership with Slashdot Media, is your hub for all
things parallel software development, from weekly thought leadership blogs to
news, videos, case studies, tutorials and more. Take a look and join the
conversation now. http://goparallel.sourceforge.net/

_______________________________________________
Nvda-devel mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/nvda-devel


--
Doug Lee                 [hidden email]                http://www.dlee.org
SSB BART Group           [hidden email]   http://www.ssbbartgroup.com
In healthy competition, the best battles are not for status, but for excellence;
and the battles are not between me and you,
but between you and you, and between me and me.  (08/15/2009)



------------------------------------------------------------------------------
Dive into the World of Parallel Programming The Go Parallel Website, sponsored
by Intel and developed in partnership with Slashdot Media, is your hub for all
things parallel software development, from weekly thought leadership blogs to
news, videos, case studies, tutorials and more. Take a look and join the
conversation now. http://goparallel.sourceforge.net/
_______________________________________________
Nvda-devel mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/nvda-devel



------------------------------------------------------------------------------
Dive into the World of Parallel Programming The Go Parallel Website, sponsored
by Intel and developed in partnership with Slashdot Media, is your hub for all
things parallel software development, from weekly thought leadership blogs to
news, videos, case studies, tutorials and more. Take a look and join the
conversation now. http://goparallel.sourceforge.net/
_______________________________________________
Nvda-devel mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/nvda-devel
Reply | Threaded
Open this post in threaded view
|

Re: Frustrated with NVDA: interferes in the behavior of the page.

Aaron Cannon
Hi Hedi.

If radio buttons are not resizing with magnification in Firefox, then
you should probably submit a bug report, as that definitely seems like
a problem.

In the meantime, you could always create a simulated radio button,
with some CSS, HTML, JavaScript and ARIA magic.

Here's a working example that might get you started:
http://oaa-accessibility.org/example/31/

Good luck.

Aaron

On 3/30/15, Hedi Carlos Minin <[hidden email]> wrote:

> Thank you guys, I'm glad to respond.
> I'm learning a lot.
> It seems I'm being boring insisting on something that will not work, sorry.
>
> The objective of this work is to identify requirements for delivery of an
> examination applied every year in Brazil for middle school students through
> the computer for the blind.
>
> After all these answers I realize that the best thing to do is create the
> complication using standard elements as much as possible and not try to
> change them, leaving the user in control.
>
> I will perform usability tests, then I can test two applications and ask
> the user which you prefer.
>
> This is my first work specifically for the visually impaired, and I realize
> that there are many doubts about many details. Every day I learn a lot.
>
> Sorry for my English, is Google translate.
>
>
> 2015-03-30 11:25 GMT-03:00 Doug Lee <[hidden email]>:
>
>> Hedi,
>>
>> James is in Australia and thus not likely to respond again for several
>> hours, if I had to guess. I would consider him an authority in here on
>> things like this. That said though...
>>
>> I have been professionally adapting screen readers to work with
>> applications for over 15 years, and I thought it might be good for me
>> to weigh in...
>>
>> A blind person gets used to speech patterns much as a sighted person
>> gets used to visual ones. When the pattern changes, in both cases, we
>> naturally wonder why, and may sometimes even be confused by the
>> change.
>>
>> So, for a list, the absence of "1 of 5," etc. might make a blind
>> person wonder whether this is really a list. Of course in context, it
>> may quickly become obvious for your situation; but initially, it may
>> be a sort of confusion that could be avoided.
>>
>> I don't presume to give you the right way to code this; I simply
>> suggest that the reactions you get from blind users on this list may
>> help you see more clearly what really could confuse us, based on what
>> we have come over many years to expect.
>>
>> I hope this is helpful, and I am very appreciative that you have come
>> to a list composed of blind users and developers to discuss how best
>> to develop something we are meant to use. It could go without saying,
>> but it won't: We would love to see more of this! :-)
>>
>> On Mon, Mar 30, 2015 at 11:13:42AM -0300, Hedi Carlos Minin wrote:
>>    Aaron, I agree with you on some points.
>>    We want to use screen readers because users are already accustomed to
>>    using and we should consider it.
>>    Navigation should be simple: a few elements and able to go through
>> each
>>    element of the application with Tab and Shift + Tab.
>>    A question has a clear format: a few paragraphs containing the
>>    description and five response options. There are 90 questions, then
>>    announce "paragraph" after each paragraph and "list item" after each
>>    response option becomes redundant, tiring and can cause confusion on
>>    some question. Only are these two elements I am avoiding the speech
>> of
>>    the role.
>>    Of course, a button must be advertised as a button, so that users know
>>    you can perform an action.
>>    The application is also for people with low vision. The problem I
>>    encountered when using standard radio controls is that it does not
>>    increase with the zoom on some browsers (like Firefox).
>>    I had considered a desktop application, but display mathml, svg, and
>>    have control over zoom, constrast, is not a trivial task. For that
>> HTML
>>    5 is an interesting option.
>>    I also considered one self voice application. In JavaFX I have a Web
>>    browser, and have control over various aspects. My problem is that not
>>    only hear the questions, you need to write well. I would have to
>>    "emulate" a screen reader, which is not an easy task.
>>    Well I think I should choose a browser (such as Firefox) and design
>> the
>>    application to be compatible with it.
>>    Hedi
>>
>>    2015-03-30 10:16 GMT-03:00 Aaron Cannon
>>    <[1][hidden email]>:
>>
>>      The big problem, in my mind, with what you're doing is you are
>>      trying
>>      to make decisions about what a user will and will not expect to
>>      hear.
>>      Screen reader users know that you are supposed to press enter to
>>      click.  Screen reader users are used to hearing lists, and should
>>      not
>>      be surprised when they are told that a list of answers is in fact
>>      one.
>>      In my experience, one of the hardest things for persons new to web
>>      accessibility is knowing when a screen reader is doing something
>>      expected or not.  Things that can seem like a bug are often not,
>>      and
>>      vice versa.
>>      Also, if you are working with Chrome, you should know that you are
>>      working with the browser, which, I believe, is the least well
>>      supported by NVDA, or pretty much any windows screen reader.  If
>>      you
>>      have a choice, I'd recommend Firefox and/or IE, with my preferred
>>      browser being FF.
>>      If I were you, unless you have some special requirements that you
>>      haven't shared, I'd just trust your users to figure it out, and just
>>      mark up your page with standard HTML controls.  Make your answers
>>      radio buttons or checkboxes (if multiple answers are permitted).
>>      Trust your user to know that they can press enter to click a link or
>>      button.  Otherwise, it seems to me that you may be attempting to
>>      reinvent your own screen reader on top of an existing one.
>>      If you really do need complete control over exactly what is spoken,
>>      then I fear your only option may be to create a self-voicing
>>      application, something which is possible but very fragile as a web
>>      application.  Most self-voicing apps are native applications.
>>      Good luck.
>>      Aaron
>>
>>    On 3/30/15, Rui Fontes <[2][hidden email]> wrote:
>>    > If it is for a specific application, why not create a profil where
>>    the
>>    > announcement of the position is disabled?
>>    >
>>    > Rui
>>    >
>>    >
>>    > De: Hedi Carlos Minin
>>    > Data: segunda-feira, 30 de Março de 2015 01:46
>>    > Para: NVDA screen reader development
>>    > Assunto: Re: [Nvda-devel] Frustrated with NVDA: interferes in the
>>    behavior
>>    > of the page.
>>    >
>>    > As this is an application for test delivery, where the candidate is
>>    being
>>    > evaluated, I need to have a minimum of speech (just enough). I do
>> not
>>    need
>>    > to announce a list nor its entirety, because the list of options
>> will
>>    always
>>    > have 5 elements (A, B, C, D and E). The announcement of the elements
>>    can
>>    > cause confusion, for example, if I have the list:
>>    >
>>    > A: 2.
>>    > B: 4.
>>    > C: 7.
>>    > D: 9
>>    > E: 11
>>    >
>>    > and NVDA reads:
>>    >
>>    > A: 2 1 of 5
>>    > B: 4 2 of 5
>>    >
>>    > C: 7 3 of 5
>>    >
>>    > D: 9 4 of 5
>>    >
>>    > E: 11 5 of 5
>>    >
>>    > The memory is very short and this additional talk can be
>> distracting.
>>    This
>>    > is a very specific application and I need to have total control over
>>    what is
>>    > or is not spoken. But I'm realizing that this is not an easy task.
>>    >
>>    >
>>    >
>>
>>      >
>>      --------------------------------------------------------------------
>>      ------------
>>      >
>>      --------------------------------------------------------------------
>>      ----------
>>      > Dive into the World of Parallel Programming The Go Parallel
>>      Website,
>>      > sponsored
>>      > by Intel and developed in partnership with Slashdot Media, is your
>>      hub for
>>      > all
>>      > things parallel software development, from weekly thought
>>      leadership blogs
>>      > to
>>      > news, videos, case studies, tutorials and more. Take a look and
>>      join the
>>      > conversation now. [3]http://goparallel.sourceforge.net/
>>      >
>>      >
>>      >
>>      --------------------------------------------------------------------
>>      ------------
>>      > _______________________________________________
>>      > Nvda-devel mailing list
>>      > [4][hidden email]
>>      > [5]https://lists.sourceforge.net/lists/listinfo/nvda-devel
>>      >
>>      --------------------------------------------------------------------
>>      ----------
>>      Dive into the World of Parallel Programming The Go Parallel Website,
>>      sponsored
>>      by Intel and developed in partnership with Slashdot Media, is your
>>      hub for all
>>      things parallel software development, from weekly thought leadership
>>      blogs to
>>      news, videos, case studies, tutorials and more. Take a look and join
>>      the
>>      conversation now. [6]http://goparallel.sourceforge.net/
>>
>>    _______________________________________________
>>    Nvda-devel mailing list
>>    [7][hidden email]
>>    [8]https://lists.sourceforge.net/lists/listinfo/nvda-devel
>>
>> References
>>
>>    1. mailto:[hidden email]
>>    2. mailto:[hidden email]
>>    3. http://goparallel.sourceforge.net/
>>    4. mailto:[hidden email]
>>    5. https://lists.sourceforge.net/lists/listinfo/nvda-devel
>>    6. http://goparallel.sourceforge.net/
>>    7. mailto:[hidden email]
>>    8. https://lists.sourceforge.net/lists/listinfo/nvda-devel
>>
>>
>> ------------------------------------------------------------------------------
>> Dive into the World of Parallel Programming The Go Parallel Website,
>> sponsored
>> by Intel and developed in partnership with Slashdot Media, is your hub
>> for
>> all
>> things parallel software development, from weekly thought leadership
>> blogs
>> to
>> news, videos, case studies, tutorials and more. Take a look and join the
>> conversation now. http://goparallel.sourceforge.net/
>>
>> _______________________________________________
>> Nvda-devel mailing list
>> [hidden email]
>> https://lists.sourceforge.net/lists/listinfo/nvda-devel
>>
>>
>> --
>> Doug Lee                 [hidden email]                http://www.dlee.org
>> SSB BART Group           [hidden email]
>> http://www.ssbbartgroup.com
>> In healthy competition, the best battles are not for status, but for
>> excellence;
>> and the battles are not between me and you,
>> but between you and you, and between me and me.  (08/15/2009)
>>
>>
>>
>>
>> ------------------------------------------------------------------------------
>> Dive into the World of Parallel Programming The Go Parallel Website,
>> sponsored
>> by Intel and developed in partnership with Slashdot Media, is your hub
>> for
>> all
>> things parallel software development, from weekly thought leadership
>> blogs
>> to
>> news, videos, case studies, tutorials and more. Take a look and join the
>> conversation now. http://goparallel.sourceforge.net/
>> _______________________________________________
>> Nvda-devel mailing list
>> [hidden email]
>> https://lists.sourceforge.net/lists/listinfo/nvda-devel
>>
>>
>

------------------------------------------------------------------------------
Dive into the World of Parallel Programming The Go Parallel Website, sponsored
by Intel and developed in partnership with Slashdot Media, is your hub for all
things parallel software development, from weekly thought leadership blogs to
news, videos, case studies, tutorials and more. Take a look and join the
conversation now. http://goparallel.sourceforge.net/
_______________________________________________
Nvda-devel mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/nvda-devel
Reply | Threaded
Open this post in threaded view
|

Re: Frustrated with NVDA: interferes in the behavior of the page.

derek riemer
In reply to this post by Hedi Carlos Minin
You should also think of just making each heading a question. This is quite useful for many such applications and it is what blind people might find easiest. Many people will be better and quicker using browse mode to navigate the questions as well.

On 3/30/2015 8:13 AM, Hedi Carlos Minin wrote:
Aaron, I agree with you on some points.
We want to use screen readers because users are already accustomed to using and we should consider it.
Navigation should be simple: a few elements and able to go through each element of the application with Tab and Shift + Tab.

A question has a clear format: a few paragraphs containing the description and five response options. There are 90 questions, then announce "paragraph" after each paragraph and "list item" after each response option becomes redundant, tiring and can cause confusion on some question. Only are these two elements I am avoiding the speech of the role.
Of course, a button must be advertised as a button, so that users know you can perform an action.

The application is also for people with low vision. The problem I encountered when using standard radio controls is that it does not increase with the zoom on some browsers (like Firefox).

I had considered a desktop application, but display mathml, svg, and have control over zoom, constrast, is not a trivial task. For that HTML 5 is an interesting option.

I also considered one self voice application. In JavaFX I have a Web browser, and have control over various aspects. My problem is that not only hear the questions, you need to write well. I would have to "emulate" a screen reader, which is not an easy task.

Well I think I should choose a browser (such as Firefox) and design the application to be compatible with it.

Hedi

2015-03-30 10:16 GMT-03:00 Aaron Cannon <[hidden email]>:
The big problem, in my mind, with what you're doing is you are trying
to make decisions about what a user will and will not expect to hear.
Screen reader users know that you are supposed to press enter to
click.  Screen reader users are used to hearing lists, and should not
be surprised when they are told that a list of answers is in fact one.

In my experience, one of the hardest things for persons new to web
accessibility is knowing when a screen reader is doing something
expected or not.  Things that can seem like a bug are often not, and
vice versa.

Also, if you are working with Chrome, you should know that you are
working with the browser, which, I believe, is the least well
supported by NVDA, or pretty much any windows screen reader.  If you
have a choice, I'd recommend Firefox and/or IE, with my preferred
browser being FF.

If I were you, unless you have some special requirements that you
haven't shared, I'd just trust your users to figure it out, and just
mark up your page with standard HTML controls.  Make your answers
radio buttons or checkboxes (if multiple answers are permitted).
Trust your user to know that they can press enter to click a link or
button.  Otherwise, it seems to me that you may be attempting to
reinvent your own screen reader on top of an existing one.

If you really do need complete control over exactly what is spoken,
then I fear your only option may be to create a self-voicing
application, something which is possible but very fragile as a web
application.  Most self-voicing apps are native applications.

Good luck.

Aaron

On 3/30/15, Rui Fontes <[hidden email]> wrote:
> If it is for a specific application, why not create a profil where the
> announcement of the position is disabled?
>
> Rui
>
>
> De: Hedi Carlos Minin
> Data: segunda-feira, 30 de Março de 2015 01:46
> Para: NVDA screen reader development
> Assunto: Re: [Nvda-devel] Frustrated with NVDA: interferes in the behavior
> of the page.
>
> As this is an application for test delivery, where the candidate is being
> evaluated, I need to have a minimum of speech (just enough). I do not need
> to announce a list nor its entirety, because the list of options will always
> have 5 elements (A, B, C, D and E). The announcement of the elements can
> cause confusion, for example, if I have the list:
>
> A: 2.
> B: 4.
> C: 7.
> D: 9
> E: 11
>
> and NVDA reads:
>
> A: 2 1 of 5
> B: 4 2 of 5
>
> C: 7 3 of 5
>
> D: 9 4 of 5
>
> E: 11 5 of 5
>
> The memory is very short and this additional talk can be distracting. This
> is a very specific application and I need to have total control over what is
> or is not spoken. But I'm realizing that this is not an easy task.
>
>
>
> --------------------------------------------------------------------------------
> ------------------------------------------------------------------------------
> Dive into the World of Parallel Programming The Go Parallel Website,
> sponsored
> by Intel and developed in partnership with Slashdot Media, is your hub for
> all
> things parallel software development, from weekly thought leadership blogs
> to
> news, videos, case studies, tutorials and more. Take a look and join the
> conversation now. http://goparallel.sourceforge.net/
>
>
> --------------------------------------------------------------------------------
> _______________________________________________
> Nvda-devel mailing list
> [hidden email]
> https://lists.sourceforge.net/lists/listinfo/nvda-devel
>

------------------------------------------------------------------------------
Dive into the World of Parallel Programming The Go Parallel Website, sponsored
by Intel and developed in partnership with Slashdot Media, is your hub for all
things parallel software development, from weekly thought leadership blogs to
news, videos, case studies, tutorials and more. Take a look and join the
conversation now. http://goparallel.sourceforge.net/
_______________________________________________
Nvda-devel mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/nvda-devel



------------------------------------------------------------------------------
Dive into the World of Parallel Programming The Go Parallel Website, sponsored
by Intel and developed in partnership with Slashdot Media, is your hub for all
things parallel software development, from weekly thought leadership blogs to
news, videos, case studies, tutorials and more. Take a look and join the 
conversation now. http://goparallel.sourceforge.net/


_______________________________________________
Nvda-devel mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/nvda-devel


------------------------------------------------------------------------------
Dive into the World of Parallel Programming The Go Parallel Website, sponsored
by Intel and developed in partnership with Slashdot Media, is your hub for all
things parallel software development, from weekly thought leadership blogs to
news, videos, case studies, tutorials and more. Take a look and join the
conversation now. http://goparallel.sourceforge.net/
_______________________________________________
Nvda-devel mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/nvda-devel
Reply | Threaded
Open this post in threaded view
|

Re: Frustrated with NVDA: interferes in the behavior of the page.

Brian Gaff Lineone downstairs
Though I may be speaking a little out of turn here. the problem with using
just one browser is that then you have to make people aware their browser if
its something else, cannot be used or you will get an inbox full of
complaints.
 Now of course I have not seen your  form, but I have found good an bad on
the web, and in my opinion, the very simplest ar the best. Even if they have
to rely on extra instructions for screenreader users to read the extra text
you need them to read.
 Using very modern versions of html is going to break a lot of folks systems
using older browsers and screenreaders. it obviously depends on your
potential users abilities and hard/software as to which way you go on such
things.
 Brian

[hidden email]
Brian Gaff's other account.

----- Original Message -----
From: "derek riemer" <[hidden email]>
To: "NVDA screen reader development" <[hidden email]>
Sent: Wednesday, April 01, 2015 6:25 AM
Subject: Re: [Nvda-devel] Frustrated with NVDA: interferes in the behavior
of the page.


> You should also think of just making each heading a question. This is
> quite useful for many such applications and it is what blind people
> might find easiest. Many people will be better and quicker using browse
> mode to navigate the questions as well.
>
> On 3/30/2015 8:13 AM, Hedi Carlos Minin wrote:
>> Aaron, I agree with you on some points.
>> We want to use screen readers because users are already accustomed to
>> using and we should consider it.
>> Navigation should be simple: a few elements and able to go through
>> each element of the application with Tab and Shift + Tab.
>>
>> A question has a clear format: a few paragraphs containing the
>> description and five response options. There are 90 questions, then
>> announce "paragraph" after each paragraph and "list item" after each
>> response option becomes redundant, tiring and can cause confusion on
>> some question. Only are these two elements I am avoiding the speech of
>> the role.
>> Of course, a button must be advertised as a button, so that users know
>> you can perform an action.
>>
>> The application is also for people with low vision. The problem I
>> encountered when using standard radio controls is that it does not
>> increase with the zoom on some browsers (like Firefox).
>>
>> I had considered a desktop application, but display mathml, svg, and
>> have control over zoom, constrast, is not a trivial task. For that
>> HTML 5 is an interesting option.
>>
>> I also considered one self voice application. In JavaFX I have a Web
>> browser, and have control over various aspects. My problem is that not
>> only hear the questions, you need to write well. I would have to
>> "emulate" a screen reader, which is not an easy task.
>>
>> Well I think I should choose a browser (such as Firefox) and design
>> the application to be compatible with it.
>>
>> Hedi
>>
>> 2015-03-30 10:16 GMT-03:00 Aaron Cannon
>> <[hidden email] <mailto:[hidden email]>>:
>>
>>     The big problem, in my mind, with what you're doing is you are trying
>>     to make decisions about what a user will and will not expect to hear.
>>     Screen reader users know that you are supposed to press enter to
>>     click.  Screen reader users are used to hearing lists, and should not
>>     be surprised when they are told that a list of answers is in fact
>> one.
>>
>>     In my experience, one of the hardest things for persons new to web
>>     accessibility is knowing when a screen reader is doing something
>>     expected or not.  Things that can seem like a bug are often not, and
>>     vice versa.
>>
>>     Also, if you are working with Chrome, you should know that you are
>>     working with the browser, which, I believe, is the least well
>>     supported by NVDA, or pretty much any windows screen reader.  If you
>>     have a choice, I'd recommend Firefox and/or IE, with my preferred
>>     browser being FF.
>>
>>     If I were you, unless you have some special requirements that you
>>     haven't shared, I'd just trust your users to figure it out, and just
>>     mark up your page with standard HTML controls.  Make your answers
>>     radio buttons or checkboxes (if multiple answers are permitted).
>>     Trust your user to know that they can press enter to click a link or
>>     button.  Otherwise, it seems to me that you may be attempting to
>>     reinvent your own screen reader on top of an existing one.
>>
>>     If you really do need complete control over exactly what is spoken,
>>     then I fear your only option may be to create a self-voicing
>>     application, something which is possible but very fragile as a web
>>     application.  Most self-voicing apps are native applications.
>>
>>     Good luck.
>>
>>     Aaron
>>
>>     On 3/30/15, Rui Fontes <[hidden email]
>>     <mailto:[hidden email]>> wrote:
>>     > If it is for a specific application, why not create a profil
>>     where the
>>     > announcement of the position is disabled?
>>     >
>>     > Rui
>>     >
>>     >
>>     > De: Hedi Carlos Minin
>>     > Data: segunda-feira, 30 de Março de 2015 01:46
>>     > Para: NVDA screen reader development
>>     > Assunto: Re: [Nvda-devel] Frustrated with NVDA: interferes in
>>     the behavior
>>     > of the page.
>>     >
>>     > As this is an application for test delivery, where the candidate
>>     is being
>>     > evaluated, I need to have a minimum of speech (just enough). I
>>     do not need
>>     > to announce a list nor its entirety, because the list of options
>>     will always
>>     > have 5 elements (A, B, C, D and E). The announcement of the
>>     elements can
>>     > cause confusion, for example, if I have the list:
>>     >
>>     > A: 2.
>>     > B: 4.
>>     > C: 7.
>>     > D: 9
>>     > E: 11
>>     >
>>     > and NVDA reads:
>>     >
>>     > A: 2 1 of 5
>>     > B: 4 2 of 5
>>     >
>>     > C: 7 3 of 5
>>     >
>>     > D: 9 4 of 5
>>     >
>>     > E: 11 5 of 5
>>     >
>>     > The memory is very short and this additional talk can be
>>     distracting. This
>>     > is a very specific application and I need to have total control
>>     over what is
>>     > or is not spoken. But I'm realizing that this is not an easy task.
>>     >
>>     >
>>     >
>>     >
>>     --------------------------------------------------------------------------------
>>     > ------------------------------------------------------------------------------
>>     > Dive into the World of Parallel Programming The Go Parallel
>> Website,
>>     > sponsored
>>     > by Intel and developed in partnership with Slashdot Media, is
>>     your hub for
>>     > all
>>     > things parallel software development, from weekly thought
>>     leadership blogs
>>     > to
>>     > news, videos, case studies, tutorials and more. Take a look and
>>     join the
>>     > conversation now. http://goparallel.sourceforge.net/
>>     >
>>     >
>>     >
>>     --------------------------------------------------------------------------------
>>     > _______________________________________________
>>     > Nvda-devel mailing list
>>     > [hidden email]
>>     <mailto:[hidden email]>
>>     > https://lists.sourceforge.net/lists/listinfo/nvda-devel
>>     >
>>
>>     ------------------------------------------------------------------------------
>>     Dive into the World of Parallel Programming The Go Parallel
>>     Website, sponsored
>>     by Intel and developed in partnership with Slashdot Media, is your
>>     hub for all
>>     things parallel software development, from weekly thought
>>     leadership blogs to
>>     news, videos, case studies, tutorials and more. Take a look and
>>     join the
>>     conversation now. http://goparallel.sourceforge.net/
>>     _______________________________________________
>>     Nvda-devel mailing list
>>     [hidden email]
>>     <mailto:[hidden email]>
>>     https://lists.sourceforge.net/lists/listinfo/nvda-devel
>>
>>
>>
>>
>> ------------------------------------------------------------------------------
>> Dive into the World of Parallel Programming The Go Parallel Website,
>> sponsored
>> by Intel and developed in partnership with Slashdot Media, is your hub
>> for all
>> things parallel software development, from weekly thought leadership
>> blogs to
>> news, videos, case studies, tutorials and more. Take a look and join the
>> conversation now. http://goparallel.sourceforge.net/
>>
>>
>> _______________________________________________
>> Nvda-devel mailing list
>> [hidden email]
>> https://lists.sourceforge.net/lists/listinfo/nvda-devel
>
>


--------------------------------------------------------------------------------


> ------------------------------------------------------------------------------
> Dive into the World of Parallel Programming The Go Parallel Website,
> sponsored
> by Intel and developed in partnership with Slashdot Media, is your hub for
> all
> things parallel software development, from weekly thought leadership blogs
> to
> news, videos, case studies, tutorials and more. Take a look and join the
> conversation now. http://goparallel.sourceforge.net/


--------------------------------------------------------------------------------


> _______________________________________________
> Nvda-devel mailing list
> [hidden email]
> https://lists.sourceforge.net/lists/listinfo/nvda-devel
>


------------------------------------------------------------------------------
Dive into the World of Parallel Programming The Go Parallel Website, sponsored
by Intel and developed in partnership with Slashdot Media, is your hub for all
things parallel software development, from weekly thought leadership blogs to
news, videos, case studies, tutorials and more. Take a look and join the
conversation now. http://goparallel.sourceforge.net/
_______________________________________________
Nvda-devel mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/nvda-devel