W3C home > Mailing lists > Public > w3c-wai-ig@w3.org > January to March 2002

Re: Text direction, CSS and accessibility

From: Al Gilman <asgilman@iamdigex.net>
Date: Sun, 24 Feb 2002 11:56:57 -0500
Message-Id: <200202241656.LAA199975@smtp2.mail.iamworld.net>
To: Chris Croome <chris@webarchitects.co.uk>, w3c-wai-ig@w3.org
At 06:22 AM 2002-02-24 , Chris Croome wrote:
>So back to my original question -- is it the case that one should not
>use CSS for setting direction since this seems to go against the WCAG?

Since setting this property in the HTML does not require you _not to use CSS_, then, No.  If I read the guidelines correctly.  You should simply not use _CSS alone_ to accomplish this.

And you should be looking for positive steps to take to deliver your Urdu content, not a guidelines excuse for why not to do something.

Don't expect this list to provide either authoritative interpretation of the WCAG Recommendation, or reliable advice concerning how to make Urdu work on the web.  The web offers resources, such as

 Urdu language resources: Mumbai/Bombay pages (fonts, software, ...
 http://theory.tifr.res.in/bombay/history/people/language/urdu.html

that can help you do the latter.  If you have a test page you want to cross-check for accessibility problems, bring it back here when you have figured out what you think will work for your Urdu speakers.

In other words, we can now upgrade "RTFM" to "RTFW."  That's "Read The <expletive/> Manual" in the original and "Read The <expletive/> Web" in the update, for those not familiar with the news.answers newsgroup and the history of the Internet as a knowledge community [1].

The above was found just now by a Google search using "Urdu language resources" as the search phrase based on the conversation that we had earlier about Bengali.  I have no practical experience publishing Urdu or anything but English to the Web.

You may skip the rants below if you wish.

Al

-- notes

[1] RTFM.mit.edu and the evolution of the Internet as an encyclopedia:
 http://www-unix.gridforum.org/mail_archive/gce-wg/msg00134.html

People think the contest for "Who invented the Web" is between Tim BL and Marc Andreesen.  But delivering Power to the People hangs in large measure on the thread of development started by the Harvest system out of U. of Colorado, a thread of development which users today know as Google and its competing search services.  


-- rants

It appears to me that your interpretation of your "original question" may be more juridical than is appropriate for this forum.

If you want an authoritative interpretation of the WCAG 1.0 you should request of the chairs of the WCAG working group (a.k.a. GL) for an interpretation.  If you are looking for techniques that work from fellow practitioners such as David, then this is a good place to raise the question.

As in all cases where, if CSS is implemented correctly, the CSS overrides the properties set in the HTML, there is no basis, in _my reading_ of the WCAG 1.0 to say either "don't use CSS" or "don't use HTML" to set this property.  If you set it redundantly to the same value in both places I would argue with the GL group that you should be recognized to be in compliance.

The guideline that you point out suggests that you should definitely not leave out the markings in the HTML [or XML or XHTML etc.].  This does not mean you should not also assert the correct direction in CSS rules if you find that some implementations are sensitive to one and some to the other.

Again, I would argue that WCAG 1.0 was not field tested in Urdu-speaking communities and as a result, should itself to be taken with a grain of salt when rendering services with Urdu-language web content.

For languages that are not already prevalent on the Web, as for obscure disabilities such as deaf-blind quadraplegics, you will probably find that the guidelines are not enough to guarantee success and you have to do some targeted research to find a pattern of practice that works for your users.

>
>Then why does the HTML4 spec have a whole page dedicated to how the dir
>attribute should be used [1], why is the a big section in CSS2 on this
>matter [2] and why are the Unicode Consortium and the W3C producing
>notes with sections of the use of the dir attribute [3]? 
>

There is that much expository "voice over" precisely because you can't assume support in the current operating practices.

The Unicode consortium and the W3C are engaged in creating a technical specification by means of which all these language groups can operate through computers and the Web with the maximum use of common technology across languages.  This kind of omnibus utility is not necessarily what minority-language support is first implemented with.  

>Hi
>
>On Sun 24-Feb-2002 at 09:13:53 +0000, David Woolley wrote:
>> > Therefore would I be correct in deducting that there is no point in

>> > setting text direction using CSS [2] since it has to be set in the
>> > document using the dir attribute in order to be accessible?
>> 
>> My understanding is that it should very rarely need to be set in the
>> document as the use of Arabic Unicode characters should automatically
>> trigger the relevant direction change.
>
>Then why does the HTML4 spec have a whole page dedicated to how the dir
>attribute should be used [1], why is the a big section in CSS2 on this
>matter [2] and why are the Unicode Consortium and the W3C producing
>notes with sections of the use of the dir attribute [3]? 
>
>> I'd caution, though, that few browsers can actually cope with right to
>> left text and you might need to give specific overrides if you start
>> with text as images.  
>
>I'm not planning to use any text for images (apart from logos etc).
>
>> Even IE doensn't support the style of Arabic script normally used for
>> Urdu, as far as I can recall; at one stage, this had to be hand
>> written in newspapers even though the other form could be typeset or
>> typewritten.
>
>I've not come across any Unicode Urdu myself yet, but the Hebrew [4] and
>Persian [5] pages on the Unicode site seem to work fine in Windows and
>Linux (and these pages _do_ specify the direction of the text).
>
>In mozilla the Hewbrew page has a H1 that, visually, _starts_ with 
>"(Unicode)? ... ", if you edit the page locally and remove dir="rtl", then
>open it you get the H1 _ending_ with " ... (Unicode)?".
>
>This indicates to me that the direction attribute _is_ needed in the
>current generation of browsers and also that if the direction was set
>using CSS that this would go against the WAI WCAG Priority 1 guideline
>that documents should be accessible _without_ CSS [6].
>
>So back to my original question -- is it the case that one should not
>use CSS for setting direction since this seems to go against the WCAG?
>
>Chris
>
>
>[1] http://www.w3.org/TR/html4/struct/dirlang.html
>
>[2] http://www.w3.org/TR/REC-CSS2/visuren.html#direction
>
>[3] http://www.w3.org/TR/2002/NOTE-unicode-xml-20020218/#Bidi
>
>[4] http://www.unicode.org/unicode/standard/translations/hebrew.html
>
>[5] http://www.unicode.org/unicode/standard/translations/persian.html
>
>[6] http://www.w3.org/TR/WCAG10/wai-pageauth.html#tech-order-style-sheets
>
>-- 
>Chris Croome                               <chris@webarchitects.co.uk>
>web design                             http://www.webarchitects.co.uk/ 
>web content management                               http://mkdoc.com/   
>everything else                               http://chris.croome.net/  
> 
Received on Sunday, 24 February 2002 11:57:02 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Tuesday, 19 July 2011 18:14:00 GMT