- From: Florian Rivoal <florianr@opera.com>
- Date: Fri, 15 Jul 2011 19:28:00 +0900
- To: "Ambrose LI" <ambrose.li@gmail.com>, "Koji Ishii" <kojiishi@gluesoft.co.jp>
- Cc: "www-style@w3.org" <www-style@w3.org>
> However, text-orientation is a separate issue. We're talking about > orientation of ASCII characters within Japanese text for instance. Is it > Japanese or is it English? Do authors have to wrap all such characters > with spans and apply lang="en" to display properly? > > I still believe the way current ED is trying to solve this issue -- > 'text-orientation: vertical-right' doing what East Asian authors expect, > and 'sideways' doing what Latin authors expect -- is the right way to go. Compare these the 2 runs of numbers in this string. 松本さんが1970年に生まれたので、英語で「I was born in 1970」と自己紹介したらいいです。 I you use vertical-right to layout this, I think East Asian authors would expect the first 1970 to be upright, and the second one to be sideways. If the rule is simply to display that ascii characters sideways in vertical-right, we don't get that result, unless you require the first 1970 to be written in with fullwidth characters (and I don't think we should require that, if it can be avoided). Korean also uses the ascii period ".", not "。" to end sentences, so you can't just turn acsii characters sideways, and get what East Asian authors want. With that in mind, I think we need something like Fantasai's bidi-resembling proposal (even if it is not expressed in terms of bidi), to determine the orientation of neutral characters based on surrounding characters and the base paragraph orientation, and the lang attribute of the various tags. - Florian
Received on Friday, 15 July 2011 10:28:46 UTC