- From: Martin J. Dürst <duerst@it.aoyama.ac.jp>
- Date: Tue, 27 Sep 2016 12:06:15 +0900
- To: fantasai <fantasai.lists@inkedblade.net>, <www-international@w3.org>
On 2016/09/27 11:35, fantasai wrote: > On 09/26/2016 03:22 AM, Martin J. Dürst wrote: >> >> 1) Arabic or hebrew script displayed with a font that uses equivalent >> Latin glyphs, and for which the direction has to be >> fixed because RTL Latin doesn't make sense. > > I think at this point your document is in fact a transcription and > should be encoded as such. :) > We have the -Latn subtag for a reason. Well, if I'm not good at reading Arabic or Hebrew, I might make this up as a user stylesheet. Or I might offer this from the server without wanting to create multiple documents. >> 2) XML or XML-like documents where direction is given by an attribute, >> but this attribute (or attributes) are not named "dir", >> and/or their values are not named "rtl", "ltr",... >> >> 3) XML or XML-like documents where certain fields (appearing scattered >> throughout the document) are RTL by default or by >> definition, and where adding a dir='rtl' attribute on each of them >> would be overkill. >> >> So the summary is "You shouldn't use this unless you're an UA or >> you're really exactly sure of what you're doing." > > These are both cases where you're pretending to be a UA, effectively, > by interpreting the markup yourself where the UA is lacking. Yes, because the UA cannot know any and all kinds of markup in advance. > Under no circumstances do these situations apply to HTML documents, > which have their own 'dir' attribute for which UAs already have support. For 2) and 3), I agree. But is CSS these days limited to HTML? Regards, Martin.
Received on Tuesday, 27 September 2016 03:06:59 UTC