W3C home > Mailing lists > Public > www-style@w3.org > November 2014

Re: [css-ruby] spanning of ruby annotations across excess bases

From: Koji Ishii <kojiishi@gmail.com>
Date: Wed, 12 Nov 2014 23:34:43 +0900
Message-ID: <CAN9ydbXic_OowTpYQ39hViLSN7VnhD+145xb=Sc0RnUc9Npghg@mail.gmail.com>
To: Xidorn Quan <quanxunzhen@gmail.com>
Cc: fantasai <fantasai.lists@inkedblade.net>, kawabata taichi <kawabata.taichi@gmail.com>, "L. David Baron" <dbaron@dbaron.org>, "www-style@w3.org" <www-style@w3.org>
Sorry I guess I was a bit late to respond, but there are use cases for
spanning; it's not for an error handling. How often/rare is hard to
explain, less often than normal ruby, but probably more than one might

One thing that may help this discussion is that, spanning and line breaking
within a ruby element (mostly for light novels[1]) are for different use
cases, and they're almost never used together, so it's ok to disable line
breaking when spanning occurs. Does that help complexity?

If not, well, it's apparently better to have Ruby without spanning than not
to have, or than to wait for months to years, but it'd be the best if we
could find a way to help complexities without losing spanning.

Another note regarding JLREQ is that, it was said repeatedly so most people
here might know but, in the policy h of 1.3[2], JLREQ states that its
coverage is limited to "common books". The term "common books" is not
clearly defined, but you could read it as "novels". In such "common books",
not only spanning but also double sided ruby is really really rare, but in
other markets such as educational materials, double sided ruby is a
required feature. Another example is use of double quotes; JLREQ states one
way and the other way is completely wrong and rare, while that the other
way is the rule for other markets such as closed captions or often used in

What I wanted to say here is that, when you want to determine if a feature
is common or rare, you should think about which markets you want to target
to, and depends on that, sometimes JLREQ may not help to determine that.
The primary purpose of JLREQ is to define rules used in Japanese "common
books", not to provide data of how common a feature is used in the whole
Japanese documents.

[1] http://en.wikipedia.org/wiki/Light_novel


On Wed, Nov 12, 2014 at 12:19 PM, Xidorn Quan <quanxunzhen@gmail.com> wrote:

> On Wed, Nov 12, 2014 at 1:16 PM, fantasai <fantasai.lists@inkedblade.net>
> wrote:
>> On 11/11/2014 09:01 PM, Xidorn Quan wrote:
>>> I don't agree with that model as well, but I don't think HTML5 model is
>>> perfect either. Back to spanning, IMHO, it introduces much complexity
>>> with nearly no benefits, hence I suggest that we should get rid of it.
>> The benefit is consistency of the markup model for double-sided ruby,
>> regardless of whether you need spanning or not. See explanation at:
>>   http://fantasai.inkedblade.net/weblog/2011/ruby/
>> Without it, you'd need to switch between nested ruby or <rtc> markup
>> depending on whether that particular word in the second level required
>> spanning or broke at the same boundary as the first level.
> Do we really have use cases that authors may sometimes need span but
> sometimes not in the same level?
> AFAICS, we have two main use cases for double-sided ruby: one is the
> Japanese two pronouncation case, the other is making annotation in other
> language. For the first case, kun'yomi always spans the kanjis, while
> on'yomi are always separate. For the second case, I don't think most
> languages can be aligned with each other in sub-word level.
> Hence, it won't cause any inconsistency to require nested markup for
> spanning. Whether a level needs spanning or not doesn't change among words.
> - Xidorn
Received on Friday, 14 November 2014 08:40:06 UTC

This archive was generated by hypermail 2.3.1 : Monday, 2 May 2016 14:39:26 UTC