Re: [css3-lists] CJK numbering algorithms

On Sat, Apr 23, 2011 at 1:08 AM, Koji Ishii <> wrote:
>>>> * The fallback is used only for Korean, so I can't speak
>>>> for them much, but I guess ''cjk-decimal'' might work better.
>>>> Glyphs are closer, and the behavior in vertical writing is more
>>>> similar.
>>> Ooh, good idea.  I'll switch it.
>> Not so fast :-).  It's not a good idea to use that fallback for
>> korean-hangul-* although it makes sense to use it for korean-hanja-*.
> It sounds like this is yet another example where CJ and K are different. Maybe we should define fallback per counter style, rather than CJK as a group? I noticed that the fallback can be used for large numbers as well.

Yes, I need to rewrite the section (again) to explicitly say that the
CJK types have descriptor values, the same as if they were specified
with @counter-style.  Yesterday, I added a new 'type' value - override
- that lets you take an existing counter style and change certain
aspects, like the suffix or the range.  (By the way, Koji, this will
allow you to change the suffix without using 'content', and change the
negative sign as well.)

> I agree that not many people would complain. The same applies to Japanese. But if some may complain, why shouldn't we spend efforts to make everyone happy? If one would ever tried to use this and saw:
>  ..., -3, -2, -1, 〇, 일, 이, 삼, ...
> I guess this looks strange to you, as strange as "..., -3, -2, -1, 〇, 壱, 弐, 参, ..." looks to me.

Note that it wouldn't look like that - it would be using cjk-decimal
(or korean-decimal or whatever) for the negative numbers.  It would
still be a change in the display, but perhaps it would be less


Received on Saturday, 23 April 2011 22:22:18 UTC