- From: Xidorn Quan <quanxunzhen@gmail.com>
- Date: Thu, 12 Mar 2015 09:27:19 +1100
- To: Richard Ishida <ishida@w3.org>
- Cc: www-style list <www-style@w3.org>
- Message-ID: <CAMdq69-naWU2ZRTvamzQMdyq_EsLptKf7xNvcVHT_iLOGf_aTw@mail.gmail.com>
On Thu, Mar 12, 2015 at 5:32 AM, Richard Ishida <ishida@w3.org> wrote: > On 09/03/2015 09:48, Xidorn Quan wrote: > >> I propose to remove inter-character keyword from ruby-position, and >> handle annotation as inter-base annotation if the writing-mode of the >> text container is orthogonal to that of the base container or ruby >> container. >> >> The reasons are: >> 1. We have met some style computation problems for the keyword >> inter-character, like the cyclic dependency between writing-mode and >> display. And solving those problems seem to be non-trivial. >> 2. The behavior of an orthogonal text container is never defeined. >> 3. The keyword inter-character could be confusing, because what it does >> is not essentially inter-character. It is inter-base for horizontal >> line, and not inter- anything at all for vertical line. >> >> For these reasons, I propose this change. >> >> The only concern for this proposal is that, it will make authors have to >> add <rtc> for inter-character annotation, instead of just specifying the >> proper ruby-position on some ancestor. But I guess it could be solved if >> we backport the writing-mode of an anonymous text container from its >> first child. Backporting a property to an anonymous box might probably >> be less tricky. >> > > I may be missing the point somewhere, but these are my first reactions, > coming at this from a user perspective, rather than an immplementation > perspective. > > I think we need to keep things as simple as possible for the content > author – there are already quite a few complications for them to handle > when dealing with writing modes, and bopomofo already requires a fair > amount of markup. > > I always expected use of the 'inter-character' property to be *the thing > that triggers* vertical alignment of the ruby text (per the spec), and that > we would thereby spare the content author the trouble of adding an extra > rule for writing mode, especially given that all 'inter-character' bopomofo > is vertical anyway. > > Since this orientation is really only practically used for bopomofo ruby, > i originally suggested that the property actually be called 'bopomofo'. > Subsequent spec editors tried to make it sound a little more generic with > 'inter-character', but, i'm not sure that there are any real use cases > other than the bopmofo one (which is also what the spec implies, per my > reading). > > I'd be really disappointed if authors had to add more markup, such as > <rtc>, to make bopomofo work. Given that bopomofo *is* normally used after > every character, rather than groups of characters, it would really mess up > what was intended to be simple and easy to read/write semantic markup. Yes, I realized that. That's why I propose to make anonymous ruby text containers have their writing-mode from their first child instead of their parent, so that the authors only need rt { writing-mode: vertical-rl; } would be enough, if they don't use any <rtc>. Actually, I guess it might be good to always have anonymous ruby text containers "inherit" styles from their child instead of their parent. I'd like to open another thread to discuss this. - Xidorn
Received on Wednesday, 11 March 2015 22:28:53 UTC