- From: Richard Ishida <ishida@w3.org>
- Date: Fri, 19 Sep 2014 17:28:15 +0100
- To: Bobby Tung <bobbytung@wanderer.tw>, Koji Ishii <kojiishi@gluesoft.co.jp>, David Hyatt <hyatt@apple.com>
- CC: www-style list <www-style@w3.org>, Xiaoqian Wu <xiaoqian@w3.org>, 奕钧 陈 <ethantw@me.com>, "font@but.tw" <font@but.tw>
On 19/09/2014 15:59, Bobby Tung wrote: > There's a document published by Institute for Information Industry(iii) > Taiwan in 2011 to tell how to support EPUB 3 for Reading System[1]. In > page 52, it told that tone mark should be implied by nested ruby to make > sure be put aside. > > So I always marked Bopomofo in this way. > > <ruby>來<rt><ruby>ㄌㄞ<rt>ˊ</rt></ruby></rt></ruby> > > I don't think its an ideal way. But if we marked Bopomofo like this: > > <ruby>來<rt>ㄌㄞˊ</rt></ruby> > > To put tone mark aside may depend on 1, render engine or 2, OpenType > feature (GPOS/GSUB). > > But recently I found another issue: when Bopomobo appears inline, tone > mark as ruby[2]. > > It has same structure as nested ruby: > > ...不應該讀「<ruby>ㄧㄤ<rt>ˇ</rt></ruby>」$)A#,... > > So should Bopomofo ruby be nested or not? I am a little confusing. > > [1]: > https://www.dropbox.com/s/o68qmp707rc69jp/Publication%20software%20specification.pdf > [2]: http://lists.w3.org/Archives/Public/www-style/2014Sep/0186.html I guess two things worry me about the nested approach: [1] it's quite a long of markup - must be a pain for authors [2] it's not clear to me how the tone mark would end up in the right place, since it's not supposed to be centred, nor is it supposed to be aligned to the edge of the box according to http://www.edu.tw/files/site_content/M0001/juyin/jb99.htm?open. If it's achievable using things like CSS margins/padding, it again must be a little complicated for authors to get it right. I'm inclined to think that all this should be really simple for the content author. They should just say 'make this work like bopomofo ruby' and use a minimum of markup and CSS. I see your point about inline bopomofo also positioning tone marks in a way that looks a little like ruby, but I suspect we have the same issue in terms of positioning the tone mark correctly(?). (In fact, it may be more of an issue because the glyphs are bigger.) I also can't help feeling that it's not actually ruby, and so we ought not to use ruby markup for it. On the face of it, maybe that's a case where opentype would be useful, but I think there are also cases where the tone mark just appears as a spacing character after the zhuyin letters, no? If so, we'd need some kind of switch for content authors to indicate which they wanted. Getting back to the actual ruby case, though, is there any mileage in the idea that browsers should use opentype features if ruby tone mark positioning is supported, but otherwise use layout? That might enable us to move forward with Dave's solution in the near term and for fonts that don't support opentype, and bring in the opentype solution when/if it becomes available in the future. (I remember we already had some wording like this for tate-chu-yoko.) There's also one other aspect related to bopomofo ruby that I'm not clear about. The light tone appears at the top of the vertical stack, but there seems to be some debate about where the character U+02D9 should appear in relation to the other zhuyin letters. I believe that nominally you'd expect it to follow the characters in the zhuyin syllable, even though it is displayed before them. (It would be interesting actually, Bobby, if you have any examples of non-ruby bopomofo like the ones you just showed us that use the light tone - so we could see where it's placed - especially horizontal examples.) I noticed that https://www.moedict.tw/%E5%AD%90 stores the light tone before the zhuyin letters, though I don't know if that's just to simplify things(?). I suppose it's possible and maybe even probable that content authors would naturally type it before the syllable for ruby even if they typed it after for non-ruby zhuyin. Maybe it's not necessary to be overly purist about it. [1] (One snag may occur, I guess, if you wanted to simply replace the hanzi with the bopomofo, as you may sometimes want to do for kanji+hiragana ruby, for kids/accessibility, etc. On the other hand, I've heard that where this occurs, the bopomofo tends to still be vertical, even in horizontal text - though i'm not sure that's always the case.) ri [1] more on this at http://www.unicode.org/mail-arch/unicode-ml/y2013-m11/0053.html
Received on Friday, 19 September 2014 16:28:53 UTC