- From: Leif Halvard Silli <xn--mlform-iua@xn--mlform-iua.no>
- Date: Sun, 22 Jan 2012 22:54:56 +0100
- To: Koji Ishii <kojiishi@gluesoft.co.jp>
- Cc: Richard Ishida <ishida@w3.org>, CJK discussion <public-i18n-cjk@w3.org>
Koji Ishii, Sun, 22 Jan 2012 09:46:44 -0500: >> HTML5's letter-by-letter model creates a difference between what >> the user sees in his Web browser and the underlying code. ... snip ... > Letter-by-letter is > another example. This lack of capability can cause real problems in > fallback, selections, finding the word, TTS, and all other services > that parses HTML such as Google Translate. > > Does this express all your concerns? For letter-by-letter, two more concerns: Firstly, Spell-checking [see discussion with Martin]. Secondly, authoring. With XHTML 1.1 ruby, then, when you create letter-by-letter ruby, you first enter an expression in the ruby base container, the <rbc>. And thereafter, you you enter the 'translation' inside the <rtc>. Thus, author concentrates on writing the entire base first, and thereafter on the ruby text 'translation'. By contrast, in the current HTML5 model, the author must keep the tongue in balance and type single letters from *two* words - base and translation - simultaneously. This can probably be quite difficult if you have long word, or if you have lots of ruby to add, or if you work directly with the code. It might also be difficult to automate for hand authoring in a WYSIWYG too [consider that for the <dl> element, then - to this day - many WYSIWYG authoring tools do not offer any effective way for adding <dt> and <dd>.] How difficult it becomes for the author to keep them apart, might also depend on how different the script of the base is from the script of the ruby text. The HTML5 model, for an author, is thus comparable to a table model where one would have to work with cells from two rows simultaneously. -- Leif Halvard Silli
Received on Sunday, 22 January 2012 21:55:29 UTC