- From: Liam R E Quin <liam@w3.org>
- Date: Tue, 20 May 2014 06:06:44 -0400
- To: James Clark <jjc@jclark.com>
- Cc: Dave Cramer <dauwhe@gmail.com>, fantasai <fantasai.lists@inkedblade.net>, "www-style@w3.org" <www-style@w3.org>
On Tue, 2014-05-20 at 15:34 +0700, James Clark wrote: [...] > I agree that's the right behaviour for the Latin script. But what about > other scripts? It seems to me the real challenge in defining drop-caps for > CSS is in handling non-Latin scripts. Main issues seem to be 1. identifying which text to make large - e.g. Hebrew and Arabic texts sometimes use "drop words", suggesting use of a :first-word selector, as Dave mentioned. Hindi uses a syllable, but rather than :first-syllable :) it might be OK to use an explicit span element for now and improve selections later. The Arabic examples I linked to from my blog have both first letter and first word. 2. identifying the "base character" to be sized - and then allowing room for diacritical marks / accents / ornamentation 3. alignment - if the body text is in Arabic or Hindi, there isn't a Greek/Latin-style baseline. On the other hand, that also means that alignment can be less precise, because the human eye detects misalignment of patterned elements along edges, not misalignment of invisible character cells :-) so the examples I've seen e.g. the Arabic examples pointed to on my blog, are nominal-baseline to nominal-baseline. Devanagari does not seem to use the hanging baseline for alignment, probably because a thick line does not connect well to a thin one. 4. accessibility - making sure the markup doesn't confuse text readers, search engines, etc; I think this is not likely to be a problem but some methods used today don't work too well (e.g. tables). 5. Mixed language/script situations. This will simply need exploration with whatever approach is taken. Using a vertical-script Japanese character as a drop initial in Italian text is unlikely to work well. For sure more research is needed (and is being actively done). Drop Initials/Words seem to be relatively new to Arabic, and are copying Western practice. So the usage is evolving fast. It might be that whatever CSS decides will end up influencing what is done, both for Arabic and for at least some other scripts. So we should define what is known, and enable authors to get it right more easily than getting it wrong, and we should leave flexibility for writing systems where the rules aren't yet fixed, and we should try and learn more about others. Whatever is drafted will get lots of review, though - it's early yet. -- Liam Quin - XML Activity Lead, W3C, http://www.w3.org/People/Quin/ Pictures from old books: http://fromoldbooks.org/ Ankh: irc.sorcery.net irc.gnome.org freenode/#xml
Received on Tuesday, 20 May 2014 10:06:48 UTC