- From: Koji Ishii <kojiishi@gluesoft.co.jp>
- Date: Thu, 9 Jun 2011 10:52:44 -0400
- To: "public-html-ig-jp@w3.org" <public-html-ig-jp@w3.org>
英語MLでfantasaiが、HTMLファイルを正しく言語でタグ付すれば、CSSのauto値は正しく設定されるべきではないか、という提言をしています。 言語に依存してCSSの挙動を変える、というのは、まだ一般には許容されていません。すべての言語に対してそのような挙動を定義できるのか、そもそもHTML製作者は言語タグなんかつけないからやっても無駄じゃないか、などの意見の方が多いです。このため今のCSSでは、「日本語だったらこれとこれとこれを設定する」ということを製作者が知っている、という前提で設計がされています。 彼女の提言は日本にプラスになると思っていますが、ご意見ございましたら、お聞かせいただけますか? 英語でメール書ける方は、ぜひ元の英語MLで返事してください。私が伝えるよりも、実際に声が上がった方がより理解が進みます。 http://lists.w3.org/Archives/Public/www-style/2011Jun/0241.html よろしくお願いいたします。 -----Original Message----- From: public-i18n-core-request@w3.org [mailto:public-i18n-core-request@w3.org] On Behalf Of fantasai Sent: Thursday, June 09, 2011 5:36 PM To: www-style@w3.org; public-i18n-core@w3.org Subject: [css3-text] [css3-linebox] typesetting defaults, auto values, and i18n So, the more I work on typesetting i18n, the more obvious it becomes to me that the defaults in CSS are all skewed towards typesetting European text. A Japanese author needs to tweak a lot of things to get their page to work right, which is a problem because they also need to know that these controls exist. Ideally, CSS would have auto defaults for such properties that Just Work. But realistically, we need to reliably know the dominant script of the text, just like the bidi algorithm needs to know the dominant directionality. There are a large class of settings that need to be tweaked on a per-language or per-typesetting-style basis. I don't think that we can realistically handle all of these. There are way too many variations, and we are certain to leave out lesser-understood language-specific typesetting requirements (which, afaict, pretty much means everything except Japanese and English). But if we know the dominant script, we could handle some fairly significant things with that. For example, choosing the dominant baseline, or choosing the correct orientation of ambiguous punctuation in 'vertical-right' text. There are some 'auto' values in older drafts of our i18n-sensitive specs that are supposed to magically do the right thing, but for which the magic is undefined. We might consider having them key off the dominant writing system in documents for which this information is provided. It won't handle the language-specific variations, and we'll still need to choose a default behavior for the vast majority of untagged pages, but we could do some amount of automatic adjustment for authors who opt in by doing the extra work of tagging their documents with this information. (Authors can tag their documents by using the script subtag of the language tag.) Thoughts? ~fantasai
Received on Thursday, 9 June 2011 14:52:45 UTC