Re: [css3-lists] maintainance of counter styles WAS: Khmer consonants as list markers

On Fri, Jul 13, 2012 at 12:56 AM, Florian Rivoal <florianr@opera.com> wrote:

> On Thu, 12 Jul 2012 18:35:41 +0200, Tony Graham <tgraham@mentea.net>
> wrote:
>
>  On Thu, July 5, 2012 12:21 am, fantasai wrote:
>>
>>> On 07/04/2012 02:01 PM, Tony Graham wrote:
>>>
>> ...
>>
>>> My understanding is that an Khmer consonantal list style would be used
>>>> by
>>>> Khmer speakers if it were available to them [5].
>>>>
>>>> Can/should it be added to a future draft of css3-lists?
>>>>
>>>
>>> Seems reasonable to me. Can you put together an @counter-style rule for
>>> Tab
>>> to copy? (You'll also want to check with Khmer speakers as to what the
>>> correct
>>> behavior is once you get to the end of the alphabet.)
>>>
>>
>> Based on feedback from Sovann Heng (&#6048;&#6081;&#6020;
>> &#6047;&#6075;&#6044;&#6030;&#**6098;&#6030;) and Danh Hong, the proposed
>> @counter-style rule is:
>>
>> [...snip...]
>>
>
> I think this is a good example of why the CSS WG shouldn't take the
> responsibility for specifying counter types for that many languages, at
> least not in a normative document. There are not enough experts of all
> languages in the working group to do a good job for all languages. In this
> case, we were lucky enough to be told what we were missing, but I don't
> think this is approach is reliable enough to scale.
>

I would definitely agree. If it were a matter of simply referring to an
external normative specification for this behavior then that would be
reasonable, but it is not reasonable for CSS to attempt to create normative
orderings, which is really outside the scope of CSS. Even Unicode has
traditionally avoided defining these orderings in the past due to
controversy and lack of common agreement. As I've pointed out, in many
writing system contexts, there are more than one accepted ordering used in
modern and/or historical practice. For example, in the list proposed above,
the historical sanskrit/pali characters are left out. However, one can find
Khmer dictionaries whose ordering includes those characters.


> And if we're not sure our definitions are good or appropriate, making them
> normative and asking vendors to implement them sounds questionable to me.
>
> Maybe a better approach would be to start a community group tasked with
> developing and maintaining counter styles with a focus on i18n, and
> providing the result of their work as a stylesheet (or set of style sheets)
> that authors can use, rather than as a standard that vendors must support.
>
> This group could release updates more often than the WG could in a rec
> track document, make progress faster and achieve better coverage by
> focusing on the topic and thus potentially attract relevant experts who
> wouldn't consider joining the too generic CSS WG, provide multiple version
> simultaneously...
>
>  - Florian
>
>