W3C home > Mailing lists > Public > wai-xtech@w3.org > November 2007

Re: simple test of screen reader support for CSS-generated text

From: Gregory J. Rosmaita <oedipus@hicom.net>
Date: Tue, 20 Nov 2007 20:55:06 +0000
To: Aaron Leventhal <aaronlev@moonset.net>
Cc: wai-xtech@w3.org
Message-Id: <20071120200606.M32859@hicom.net>

aloha, aaron!

did you notice that when screen-scraping is turned ON in JAWS (that is,
when the virtual cursor is turned off), it is possible to hear the 
pseudo-content spoken, but that is NOT the default setting for JAWS when 
dealing with anything that uses a DOM or parses HTML, and it is not an 
intuitive reaction on the part of an end-user to unspoken text, because 
the non-sighted user doesn't even know that it is there, so why would 
that user turn off JAWS' virtual cursor to manually search the screen 
for the generated text, either by gross navigation or by searching for 
the generated text, which takes foreknowledge that most users will NOT 
have -- i still maintain that while reading document source is the 
course of last resort, but is an undue burden on the end user, who 
simply wants to ascertain what is on the page -- the only reason i 
tried the strategy is that it is necessary to leave virtual cursor mode 
in order to get most of the mozilla aria examples to work correctly)

curtis chen, if you're reading this, how can FireVox provide exposure
of CSS-generated content?  i do believe there are very valid use cases; 
for example, explicit contextual markers, such as "Issue: " or "Big 
Issue: " which, in the HTML5 editors' draft:

http://www.w3.org/html/wg/html5

are currently distinguished by color changes and visual markers (boxes)
alone -- moreover, the style rule defined for an "Issue: " is the same 
as that defined for a "Big Issue: ", so even if an AT could pick up and
programmatically detect the hexadecimal or rgb values defined for "issue"
and "big issue", the user still wouldn't be able to distinguish between
the two, save for the use of :before and :after to provide explicit 
markers that "this type of content starts here" and "this type of content
stops here" (yes, i know that since they are the same visually, the 
magnitude of the issue is also indistinguishable visually, this itself 
should be corrected)

as you may have guessed, this test is part of my action item to work on 
improving the accessibility of the HTML5 draft's internal stylesheet

another valid use case is for unsupported natural languages, as in the
example located at:

http://tinyurl.com/2bphmo

as well as an indication of page structure through pseudo-textual 
markers;

http://tinyurl.com/ywy66n

obviously (unless the end user wanted or needed to see them) the 
pseudo-text would be unobtrusive/not visible -- an alternate strategy 
that will work when CSS support is not available is to use the 
"overflow" technique C7 of WCAG 2.0 to visually obscure explicit 
contextual or labelling text, but keep it available to a screen 
reader and to a text-only/CSS-incapable browser, in which all styling
is lost, leaving the explicit text as contextual markers...

oh, and a question for you, aaron: does FF2 or FF3 support CSS2, 
Section 18.2, "User preferences for colors"

http://www.w3.org/TR/CSS2/ui.html#system-colors

thanks, gregory.
----------------------------------------------------------------
CONSERVATIVE, n.  A statesman who is enamored of existing evils,
as distinguished from the Liberal, who wishes to replace them 
with others.         -- Ambrose Bierce, _The Devil's Dictionary_
----------------------------------------------------------------
             Gregory J. Rosmaita, oedipus@hicom.net
  Camera Obscura: http://www.hicom.net/~oedipus/index.html
----------------------------------------------------------------

---------- Original Message -----------
From: Aaron Leventhal <aaronlev@moonset.net>
To: "Gregory J. Rosmaita" <oedipus@hicom.net>
Cc: wai-xtech@w3.org
Sent: Tue, 20 Nov 2007 12:54:16 -0500
Subject: Re: simple test of screen reader support for CSS-generated text

> Firefox 2 & 3 currently fail with a screen reader.
> 
> At the moment the only generated content we expose to 
> accessibility APIs is for list bullet text (e.g. list item 
> numbering). It's not a simple thing to do :before and :after,
>  and for the moment we've deprioritized it until we see an 
> actual use case.
> 
> - Aaron
> 
> Gregory J. Rosmaita wrote:
> > aloha!
> >
> > i have placed a very simple test of the accessibility of CSS-generated
> > content in www-archive -- the simple URI for the attachment is:
> >
> > http://tinyurl.com/2m4h2m
> >
> > explanatory text at:
> >
> > http://lists.w3.org/Archives/Public/www-archive/2007Nov/0062.html
> >
> > note that i have posted this to wai-xtech as it pertains to the work
> > of several WAI WGs
> >
> > gregory.
> > ----------------------------------------------------------------
> > CONSERVATIVE, n.  A statesman who is enamored of existing evils,
> > as distinguished from the Liberal, who wishes to replace them 
> > with others.         -- Ambrose Bierce, _The Devil's Dictionary_
> > ----------------------------------------------------------------
> >              Gregory J. Rosmaita, oedipus@hicom.net
> >   Camera Obscura: http://www.hicom.net/~oedipus/index.html
> > ----------------------------------------------------------------
------- End of Original Message -------
Received on Tuesday, 20 November 2007 20:55:41 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Friday, 27 April 2012 13:15:44 GMT