Re: [css-ruby] Ruby text and browser minimal font size

Hi Xidorn,

I'm posting at the w3c mailing list for the first time. Nice to e-meet you.
I'm jchiba, working under Hiroshi (skk@bpsinc.jp) who started to
participate in w3c discussions recently.

I noticed the same thing you mentioned while I was making a ruby sample.
A pair of Kanji with mono-ruby (two hiragana each) happened to have an
extra space
between the ruby base characters due to the browser's minimum font size.
The rendering with an extra space was not what I expected.

> I guess that user agents should probably ignore the minimum font size
> setting for ruby text elements and their descendents. And this might
> be worth a mention in the spec.

I agree.
The minimum size should not work for the ruby characters because of its
original purpose:
provide phonetic annotation using "ruby" sized typeface (smaller font size
than the body elements).

> One thing I'm concerned is that, ignoring this setting on certain
> elements effectively opens a door to authors to completely escape from
> it, which might also be unfortunate.

One suggestion to keep the door small enough is to define the minimum size
for ruby separately.
I believe 50% of the minimum body size set by the user agent is one
reasonable min value,
regarding that the current draft mentions "typically sized to 50% when used
as annotations".
It shall be 30% for bopomofo.
Or 30% for Japanese too, because somebody may want to put three hiragana
for one base without adding spaces to the base characters.

I wonder if there are other doors we should be aware of.

[1] https://drafts.csswg.org/css-ruby-1/#diagram-conventions


On Mon, 17 Aug 2015 at 15:32 Xidorn Quan <quanxunzhen@gmail.com> wrote:

> Most browsers have an option called "Minimum font size" or something
> similiar, which enables users to enforce a proper font size for them
> to read.
>
> But this feature generally doesn't work fine with ruby. If the base
> text of ruby is not larger than the minimum font size, the ruby text
> would end up having the same size with its base text. It seems to be
> unfortunate.
>
> There are some existing workarounds for authors to ensure the minimum
> font size in their page not being smaller than the user setting. But
> generally, the "minimum font size" is for the main text, not ruby. It
> doesn't make sense to enlarge text of a whole page by 1x just for
> ensuring that ruby is shown properly.
>
> I guess that user agents should probably ignore the minimum font size
> setting for ruby text elements and their descendents. And this might
> be worth a mention in the spec.
>
> One thing I'm concerned is that, ignoring this setting on certain
> elements effectively opens a door to authors to completely escape from
> it, which might also be unfortunate.
>
> Thoughts?
>
> - Xidorn
>
>