[Accessibility] Screen readers and WCAG - current state of art

Alexander Surkov surkov.alexander at gmail.com
Thu Sep 3 18:06:50 PDT 2009


Btw, I didn't find mentioned CSS techniques in WCAG2
(http://www.w3.org/TR/WCAG20-TECHS/css.html), however they were
presented in WCAG 1.0 (http://www.w3.org/TR/WCAG10-CSS-TECHS/).
Nevertheless aural CSS are presented in appendix of CSS 2.1 (though
this makes them unnecessary to be implemented by browser in order to
conform to CSS 2.1).

Alex.


On Fri, Sep 4, 2009 at 8:33 AM, Alexander
Surkov<surkov.alexander at gmail.com> wrote:
> Concerning to CSS 'speak' rule, afaik, Firefox doesn't map it to
> accessibility API. That's the reason neither NVDA nor Jaws pick it up
> I think. I assume it's true for Safary and IE since they don't work
> with them as you said.
>
> In general the problem of Aural CSS and similar innovations is they do
> not provide an implementation guide (like ARIA does for example), I
> mean, the rules how browsers should map the content to AT via AT APIs.
> For example, currently there is no way to map CSS 'speak: spell-out'
> rule to MSAA/IA2/ATK. That means browsers and screen readers should
> meet and decide how they will extend the existing API or how will they
> use existing API to adapt it to new needs. That's not quick process.
>
> The meantime development efforts are targeted on ARIA I think. ARIA is
> supported more or less well by browsers and screen readers. As well
> ARIA provides analogue of CSS 'speak: none' rule: role="presentation"
> (which is not inherited in contrast to CSS rule I guess). Thus you
> could try to use ARIA instead.
>
> The answer on "what to do?" is you should file bugs to notify
> developers and bring their attention I think.
>
> Thank you.
> Alex.
>
>
> On Fri, Sep 4, 2009 at 4:59 AM, Alexey Khoroshilov<khoroshilov at ispras.ru> wrote:
>> Hi James,
>>
>> Thank you very much for the response.
>>
>> Eugene, who is a webmaster of the site mentioned, have not got access to
>> the accessibility list yet. So he ask me to forward his answer. Please
>> find it below.
>>
>>
>> Dear Jamie,
>>
>> Thank you for your attention for Alexey’s report and fist of all I would
>> like to appreciate to the NVDA and its team.
>>
>> As for the issue with our website, WCAG and screen readers. We have two
>> ‘tricks’ with CSS on our website to make it more friendly for the
>> visually impaired visitors, e.g. hide some useless (decorative) elements
>> and on the other hand provide some additional explanations and
>> navigation elements.
>>
>> For example, we has provide to this target group additional link ‘skip
>> menu’ in the beginning of the page, etc. I guess that you would agree
>> that this link is useless for the rest of the visitors and may alter the
>> visible design of the site, so we has hide it from the visual user
>> agents (UAs).
>>
>> That was not the problem. The problem is that we couldn’t hide
>> information which is useless for screen readers and other aural UAs from
>> them.
>>
>> All of the following based on the same ‘environment’:
>> Base URL – http://www.ifap.ru/eng/index.htm
>> Host OS – Windows XP Professional (Eng/Russian MUI) Service Pack 3
>> Browser – IE 6.0.2900.5512.xpsp_sp3_gdr.090206-1234
>> Screen reader – NVDA 0.6p3.2
>>
>> The first trick is special aural CSS provided for screen readers and
>> other aural UAs as follows:
>>
>>
>> It seems that NVDA ignoring it at all and use instead ‘regular’ CSS for
>> the visual UAs provided by
>>
>> The second trick is to ‘hide’ decorative elements from aural UAs by
>> adding ‘speak: none’ in the ‘regular’ CSS for visual UAs. It also
>> doesn’t work. For example, let’s see the following code:
>>
>> HTML (index.htm):
>> some graphic with hyperlink here
>>
>>
>> CSS for visual UAs (screen.css)
>> DIV#mainlogo {speak: none}
>>
>> CSS for aural UAs (aural.css)
>> DIV#mainlogo {display: none; speak: none}
>>
>> Ooops, NVDA still speaks the content of this DIV :(
>>
>> We have slightly different behavior of the NVDA with another browsers,
>> e.g. Opera 10, Firefox 3.5.2 and Safari 4.0.3, but in that environment
>> NVDA still seems to ignore aural.css and speaks content of the mentioned
>> DIV.
>>
>> But the HTML and CSS code has passed validation through W3C Validator
>> and seems to be correct in relation to HTML 4.0/CSS 2.x/WCAG 1.0
>>
>> That was the explanation of our issue and I hope you can deal with it in
>> the future wersions of the NVDA or explain where I was wrong with the code.
>>
>> Sincerely,
>>
>> Eugene Altovsky
>> Webmaster of IFAP.ru
>>
>>
>> --
>> Alexey
>>
>>
>> James Teh wrote:
>>> Hi Alexey,
>>>
>>> I am one of NVDA's primar ydevelopers. It'd be great if you could
>>> provide more specific information about exactly what code on the web
>>> sites was not supported by the screen readers. Based on the information
>>> in the post you referenced, it would seem that the big complaint is that
>>> text that should not have been read was read and that text specifically
>>> designed for the blind was not read.
>>>
>>> To be honest, I'm not very familiar with WCAG 2.0. (I know of it, of
>>> course, but have never read it in detail.) However, I am a big believer
>>> that text designed specifically for the blind is generally a bad idea.
>>> There are of course exceptions - alt text for images, for example.
>>> However, in my opinion, the interface should be designed so as to not
>>> require a virtually "separate" interface. If the site needs to hide so
>>> much text from screen readers, this would seem to indicate a core flaw
>>> in the site's design.
>>>
>>> In any case, I'd love to hear more specific information about the
>>> problems encountered.
>>>
>>> Btw, for those that are interested, here is a link to the post using
>>> Google translate to translate it into English:
>>> http://translate.google.com/translate?prev=hp&hl=en&js=y&u=http%3A%2F%2Fcommunity.livejournal.com%2Fifap_ru%2F13520.html&sl=ru&tl=en&history_state0=
>>>
>>> Jamie
>>>
>>> On 3/09/2009 12:04 AM, Alexey Khoroshilov wrote:
>>>
>>>> Hello everyone,
>>>>
>>>> My friends from "ICO Information for All" [1] have spent significant
>>>> efforts to make their web-sites friendly to people with disabilities
>>>> according to Web Content Accessibility Guidelines 2.0 [2]. But as they
>>>> report in [3], the efforts almost had no sense, since the leading screen
>>>> readers NVDA and Jaws does not support standard instructions especially
>>>> introduced for such kind of software.
>>>>
>>>> Is it a correct understanding of the current state of art?
>>>>
>>>> [1] http://www.ifap.ru/eng/index.htm
>>>> [2] http://www.w3.org/TR/WCAG20/
>>>> [3] http://community.livejournal.com/ifap_ru/13520.html (in Russian)
>>>>
>>>> --
>>>> Alexey
>>>> _______________________________________________
>>>> Accessibility mailing list
>>>> Accessibility at lists.linux-foundation.org
>>>> https://lists.linux-foundation.org/mailman/listinfo/accessibility
>>>>
>>>
>>>
>>
>> _______________________________________________
>> Accessibility mailing list
>> Accessibility at lists.linux-foundation.org
>> https://lists.linux-foundation.org/mailman/listinfo/accessibility
>>
>


More information about the Accessibility mailing list