W3C home > Mailing lists > Public > public-css-testsuite@w3.org > April 2011

Re: bidi-004 and bidi-004a are broken

From: Alan Gresley <alan@css-class.com>
Date: Thu, 28 Apr 2011 03:14:38 +1000
Message-ID: <4DB84EFE.3090205@css-class.com>
To: fantasai <fantasai.lists@inkedblade.net>
CC: public-css-testsuite@w3.org
On 21/04/2011 3:34 AM, fantasai wrote:
> On 04/19/2011 11:48 PM, Ian Hickson wrote:
>> On Tue, 19 Apr 2011, L. David Baron wrote:
>>> On Wednesday 2011-04-20 04:28 +0000, Ian Hickson wrote:
>>>> On Tue, 19 Apr 2011, L. David Baron wrote:
>>>>>>
>>>>>> I didn't write 004a so I'm not sure what it's testing or how to fix
>>>>>> it.
>>>>>
>>>>> bidi-004a is intended to be a simplified version of bidi-004 that
>>>>> doesn't test interaction with whitespace collapsing, since we had (I
>>>>> think) no correct implementations of the interaction of bidi with
>>>>> whitespace collapsing, and had to loosen the spec to enter PR. For
>>>>> (some) more details, see:
>>>>> http://www.w3.org/mid/4D41E88E.3040808@inkedblade.net
>>>>
>>>> That seems a bit lame. What's the point of entering PR if we don't have
>>>> interop?
>>>
>>> Is there a reasonable point at which you'd expect to stop finding lack
>>> of interop?
>>
>> No. Software, specifications, and test suites are never finished, only
>> abandoned. Why abandon the spec? It's still in use. It doesn't seem
>> helpful to declare failure and stop the convergence here.
>
> We're not abandoning the spec, we're branching off a release and
> continuing trunk development. The trunk just happens to be modularized,
> which is easier for us to maintain. The requirements you're concerned
> with are all spelled out in CSS3 Text.
>
> ~fantasai


Fantasai,

I open up a text editor which code in LTR and code this markup.

   <h1>אבגדהוזח.</h1>


Now between 'הוזח' and 'אבגד'  I can insert more markup.


If I insert a <br> I get this.

   <h1>אבגד<br>הוזח.</h1>


If I insert a <span> I get this.

   <h1>אבגד<span></span>הוזח.</h1>


I know why it happens since anything that is of a LTR script rearranges 
the bidi ordering. I do not know if RTL authors use text editors which 
code in RTL and I do not know if such an editors would rearranges the 
bidi ordering similar to the above examples.

Is CSS3 text going to solve this issue or will the RTL web still be 
broken since HTML, CSS or other web languages are coded as LTR?

Do we need to develop RTL code?

<html/> .... <html>
<lmth/> .... <lmth>


This may seem bazaar (or even ludicrous) to propose such solutions but I 
see a mangled RTL web for decades to come.

It maybe be easier to add just one new element which is <bl> for 'line 
break left' and is a counterpart to <br> which has a new meaning as 
'line-break right' but still has the behavior of breaking a line on the 
right. We could even have <brs> for 'line break start' and <bre> for 
'line break end'.

I believe the best solution is to borrow something from Hebrew or Arabic 
that doesn't break bidi ordering.

   <ب>  Arabic letter for the Latin letter 'b'.

   <ב>   Hebrew letter for the Latin letter 'b'.



-- 
Alan Gresley
http://css-3d.org/
http://css-class.com/
Received on Wednesday, 27 April 2011 17:15:08 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Wednesday, 27 April 2011 17:15:19 GMT