RE: dir=auto makes no sense for descendant user-visible attributes

Hi.

To: ntounsi@emi.ac.ma
CC: aharon@google.com; duerst@it.aoyama.ac.jp; ehsan@mozilla.com; public-i18n-bidi@w3.org
From: matial@il.ibm.com
Date: Tue, 28 Feb 2012 22:47:34 +0200
Subject: Re: dir=auto makes no sense for descendant user-visible attributes

> Najib Tounsi wrote:

  >>* attribdir not specified

    >>  o There is an element, the closest ancestor or the
element itself,

    >>    who have an explicit, valid dir attribute with
a value other

    >>    than "auto" ("rtl | ltr").
attribdir have this same value.

> <end of quote>



> This may be what Aharon meant, but this is not what I have in mind. For
me, an explicit specification of dir="auto" on the element itself
> or on the closest ancestor having an explicit dir specification should
have the same effect as setting attribsdir to "auto".

I think this is fine (but I don't use bidi all that much).  
I will check the wording of the bug later today myself to make sure I understand it all; sorry I have not done so soooner.

> Shalom (Regards),  Mati
>     Bidi Architect

 >     Globalization Center Of Competency - Bidirectional
Scripts

 >      IBM Israel

       Mobile: +972 52 2554160









> From:      
 Najib Tounsi <ntounsi@emi.ac.ma>

> To:      
 "Aharon (Vladimir)
Lanin" <aharon@google.com>

> Cc:      
 Matitiahu Allouche/Israel/IBM@IBMIL,
"Martin J. Dürst" <duerst@it.aoyama.ac.jp>, Ehsan Akhgari
<ehsan@mozilla.com>, public-i18n-bidi@w3.org

> Date:      
 28/02/2012 22:05

> Subject:    
   Re: dir=auto
makes no sense for descendant user-visible attributes



>>




>> Trying  to summarize ( the idea not the wording...).
According to my

current understanding.



>> We need a way to specify the directionality of attributes when it must
>> differ from that of the element itself
Thanks Najib for summarizing this.
(Hmm, in English at least, I'd prefer to the pronouns with the original nouns here because the plural "attributes" and singular "directionality" can be confusing:
"We need a way to specify the directionality of attributes when their directionality must

differ from the directionality of the element itself."
But this is not an important correction.)  
>> (Use cases needed to support the

>> proposal). One can make usage of LRE/RLE - PDF, but not very "user
>> friendly"...



>> A new attribute, attribsdir="ltr|rtl|auto", is proposed which
would

>> determine the directionality in which the element's user-visible

>> attributes must be displayed. A very important part of this would be the

>> default value.



>> So



 >> * attribdir = ltr", the element's attributes must be displayed
to the

    user with directionality of LTR

 >> * attribdir = rtl", the element's attributes must be displayed
to the

    user with directionality of RTL

 >> * attribdir = "auto", the direction is computed independently
for each

    element's attribute.

 >> * attribdir not specified

   >>   o There is an element, the closest ancestor or the
element itself,

        who have an explicit, valid dir attribute with
a value other

        than "auto" ("rtl | ltr").
attribdir have this same value.

   >>   o There is no such element. The element's attributes
must be

        displayed to the user with directionality of
LTR (HTML default

        directionality)

ME] I am only unsure about this last case, defaulting to left-to-right, but like I said I don't use bidi that much.
Best,

--C. E. Whitehead
cewcathar@hotmail.com 

 >> * what else...





>> Right?



>> Najib



>> . . . 



>> Aharon (Vladimir) Lanin wrote:

>>> Please note that I did not include the "bell or whistle"
of the 

>>> <input> / <textarea> exception.

>>>

>>> On Tue, Feb 28, 2012 at 2:46 PM, Aharon (Vladimir) Lanin 

>>> <aharon@google.com <mailto:aharon@google.com>>
wrote:

>>>



>> -- 

>> Najib TOUNSI (tounsi at w3.org)

>> W3C Office in Morocco (http://www.w3c.org.ma/)

>> Ecole Mohammadia d'Ingénieurs, BP. 765 Agdal-RABAT Morocco

>> Mobile: +212 (0) 661 22 00 30