W3C home > Mailing lists > Public > www-international@w3.org > July to September 2005

Re: bidi discussion list was: Bidi Markup vs Unicode control characters

From: Stephen Deach <sdeach@adobe.com>
Date: Tue, 09 Aug 2005 07:44:28 -0700
To: Tex Texin <tex@xencraft.com>, Addison Phillips <addison.phillips@quest.com>
Cc: www-international@w3.org, Richard Ishida <ishida@w3.org>, Bert Bos <bert@w3.org>
Message-id: <6.1.1.1.2.20050809065706.02086ec0@mailsj-v1.corp.adobe.com>

I thought the Unicode bidi markers were provided for cases when there was 
no higher level markup, and were intended only to eliminate ambiquous cases 
where the bidi algorithm needed guidance. (Mark Davis can confirm Unicode's 
intent, I believe he is on this mailing list.)
   Assigning explicit directionality is generally only needed: 1.) to 
establish the proper starting directional-level when the first word is not 
in the primary language for the paragraph and 2.) cases where the nesting 
of languages/scripts was unclear.  --  By providing xml:lang on the outer 
object, you can infer the correct starting direction (solving case 1) and 
by properly nesting spans you can define the nesting of languages/quotes 
(solving case 2); thus it is very rare to need the html:bdo element or the 
need to assign explicit directionality via CSS or XSLT/XSL-FO.
   These directional controls are also only needed for visual presentation. 
(I recommend applying the bidi rules in the shaping process to reorder the 
glyph-stream. I do not reorder the character-stream.) For aural 
presentation, the Unicode stream is supposed to already be in 
logical/"reading" order.

   It may be a mistake for section 508 to require the ability to completely 
turn off attached style sheets, since this will render many non-(X)HTML 
documents unusable (whether the style-sheets are CSS or XSLT). It should be 
possible for the browser/reader to override aspects of the styling such as 
background, text-color, and size for readability purposes (in CSS, by 
appending a user-specified style sheet to the existing cascade). If it is 
required to be able to turn-off/ignore the attached/referenced style sheets 
(and I am forwarding this internally to check with our accessibility and 
section 508 compliance expert), then it may be necessary to add an xml:dir 
attribute so that directionality and language can both be encoded via 
markup when xml content is not xhtml.
   It is dangerous to mix Unicode bidi markers and markup bidi controls in 
a single document, since it is too easy to produce ill-formed documents. 
The normal processing of the html:bdo element or xsl-fo:bidi-override 
element is to "behave as if the appropriate LRO, RLO, RLE, or LRE marker 
was inserted before the first character of the subtended content and a PDF 
marker were inserted after the last character of the subtended content". If 
there are already markers there or there are mismatched markers in the 
subtended content, then the result is going to be incorrect.

At 2005.08.08-17:58(-0700), Tex Texin wrote:

>Well, for me it is a case of first things first- Are there bodies
>interested in collaborating on bidi best practices? If people are
>willing to do so, than we can determine the tactics, whether its email
>threads or FAQ or other.
>If GEO will host that is fine.
>
>The documents are specific, but seem to be wrong-headed. Why recommend
>markup, if the future trend is XML doesn't have it, CSS has it and
>recommends not to use it, and btw in HTML you can use it in some places
>but not others (eg attributes)?
>
>I think we need to capture, as Bert began to do in his mail very nicely
>(thanks Bert) a survey of the recommendations, the rationale behind the
>recommendations, and then an assessment as to whether the rationales are
>in fact the right ones, and then solutions based on those. It could be
>we want to suggest alternative rationales. I would like to factor in the
>capabilities of the tools available today.
>
>What seems clear to me from this discussion is that very few people
>today really understand the context of the guidelines for bidi.
>
>It is in css, but don't use it in html, but you have to use it in xml,
>but it might not be in the profile for print, but now it is supported in
>css for html...
>
>There is a joke in here somewhere about the standards being
>bidirectional...
>;-)
>
>
>Addison Phillips wrote:
> >
> > I will note that this is specifically the mission of W3C I18N GEO WG. 
> Providing FAQ information on topics such as best practices with bidi, 
> etc., are a great help to the community at large---much greater than 
> having our comments fossilized in mail list archives. I would suggest 
> that, instead of writing very long threads here (or anywhere else), the 
> next step would be to work on specific text for a FAQ. A debate about 
> that text would produce more concrete results (and be easier to gain an 
> understanding of) than having to trace threads through the archive... and 
> it would mark a more permanent reference.
> >
> > There are some specific cases (this is one) in which the I18N Core WG 
> needs to form a kind of coherent policy for W3C specifications and their 
> implementation. Non-implementation of internationalization related 
> features is a serious problem for global accessibility of the Web and the 
> I18N WGs face the "lack of implementation" argument as a justification to 
> include "less-and-less" more-and-more frequently.
> >
> > In this particular case, I note that the W3C and Unicode have 
> documented for some time the preference of markup over control codes, 
> with reasons why, in Unicode Technical Report #20, Section 3.3:
> >
> > http://www.unicode.org/reports/tr20/#Bidi
> > http://www.w3.org/TR/unicode-xml/
> >
> > As Tex notes, this may not be a comprehensive solution to every issue, 
> since it is a general document, rather than a specific one. I feel that 
> it is specific enough as a guideline to specification writers and 
> implementers of specification. Making guidelines for specific cases will 
> be easier if done in a FAQ-like way rather than trying to write a single 
> comprehensive document, in my opinion. And a summary overview would also 
> be useful in that context as well.
> >
> > Best Regards,
> >
> > ~Addison
> >
> > Addison P. Phillips
> > Globalization Architect, Quest Software
> > Chair, W3C Internationalization Core Working Group
> >
> > Internationalization is not a feature.
> > It is an architecture.
> >
> > > -----Original Message-----
> > > From: www-international-request@w3.org [mailto:www-international-
> > > request@w3.org] On Behalf Of Tex Texin
> > > Sent: 2005?8?8? 13:14
> > > To: Sandra Bostian
> > > Cc: www-international@w3.org
> > > Subject: bidi discussion list was: Bidi Markup vs Unicode control
> > > characters
> > >
> > >
> > > Sandra, yes of course. I am not suggesting CSS be made a requirement for
> > > bidi. Only that it should be acceptable and considered best practice in
> > > the right context.
> > >
> > > At the time the standard was written, I took the warning that bidi css
> > > may be ignored by user agents to be a reflection of the state of the art
> > > at the time, since most browsers were not able to support bidi with or
> > > without css.
> > >
> > > It had not occurred to me that people were interpreting it to mean that
> > > css implementers would intentionally choose not to support it or that
> > > W3C was endorsing such a view. That would be counter to goals for
> > > internationalization and making the web accessible to all.
> > >
> > > (I will have to look around, but I am hard pressed to think of another
> > > instance where the fact that a feature is optional became a
> > > recommendation to not use it. Heck, there are many required features not
> > > implemented or incompletely implemented and we don't offer the same
> > > advice.)
> > > *****
> > > A few people have remarked about their interest in the subject. Would it
> > > make sense to create a discussion list, perhaps the I18n WG would
> > > host/moderate, specific to topics of best practices with bidi on the
> > > web. I would like to see the use cases identified (eg. your gov't
> > > requirements, needs of multilingual web sites, templating, css users,
> > > etc.) and develop solutions within the existing framework, and possibly
> > > make requests for enhancements of the standards.
> > >
> > > We could carry on here, but I am not sure the overall community wants
> > > that much commentary on the subject, and also posting to a large list
> > > might intimidate some from posting.
> > > *****
> > >
> > > tex
> > >
> > > Sandra Bostian wrote:
> > > >
> > > > For US government agencies, though, this is a section 508 requirement.
> > > Your documents must be coherent without the css. And most browsers will
> > > let you turn off style sheets and use your own (often used by those with
> > > visual disabilities). This may be a case of know your requirements 
> and who
> > > your audience is, but certain sectors either can't or should be highly
> > > wary of relying solely on css.
> > > >
> > > > Thank you all for this discussion, by the way. We are looking at a new
> > > project using Arabic and this has been highly helpful in clarifying some
> > > of the issues.
> > > >
> > > > Thanks,
> > > > Sandy
> > > >
> > > > Sandy Bostian
> > > > Digital Conversion Specialist
> > > > Library of Congress
> > > > Meeting of Frontiers: http://frontiers.loc.gov
> > > > 202-707-2342
> > > > sbos@loc.gov
> > > >
> > > > >>> Tex Texin <tex@xencraft.com> 08/07/05 6:08 PM >>>
> > > >
> > > > It is not clear to me that we need to insist on the document being
> > > > coherent without css.
> > >
> > > --
> > > -------------------------------------------------------------
> > > Tex Texin   cell: +1 781 789 1898   mailto:Tex@XenCraft.com
> > > Xen Master                          http://www.i18nGuy.com
> > >
> > > XenCraft                          http://www.XenCraft.com
> > > Making e-Business Work Around the World
> > > -------------------------------------------------------------
>
>--
>-------------------------------------------------------------
>Tex Texin   cell: +1 781 789 1898   mailto:Tex@XenCraft.com
>Xen Master                          http://www.i18nGuy.com
>
>XenCraft                            http://www.XenCraft.com
>Making e-Business Work Around the World
>-------------------------------------------------------------


---Steve Deach
    sdeach@adobe.com 
Received on Tuesday, 9 August 2005 14:47:07 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Tuesday, 2 June 2009 19:17:05 GMT