W3C home > Mailing lists > Public > www-style@w3.org > February 2013

Re: [css3-fonts] language system used when language is unknown

From: John Hudson <tiro@tiro.com>
Date: Tue, 05 Feb 2013 11:00:20 -0800
Message-ID: <511156C4.8030904@tiro.com>
CC: www-style list <www-style@w3.org>
On 05/02/13 10:33 AM, John Daggett wrote:

> The issue, listed in section 6.2 [2], is whether user agents must use
> the default language system or whether they can infer based on a
> heuristic the language system in certain contexts.  For example, for
> some scripts it might be possible to infer the language system.

> I propose allowing UA's the option of using heuristics to infer the
> language system in these cases:

>    When the language of an element is unknown, user agents
>    may either use the default language system or infer one
>    based on other heuristics if appropriate.

While I don't see a problem with allowing this presuming that UA makers 
can come up with reliable heuristics, I'm having trouble imagining 
anything other than dictionary analysis coming up with an appropriate 
OTL language system tag other than <dflt>. Generally speaking, unless a 
font has been made for a specific language with minority variant 
behaviour, the <dflt> language system is always going to conform to 
majority usage of the script, hence any inference of language system 
from script is going to result in <dflt> anyway. If a font has been made 
for minority use, then this is going to result in an incorrect inference 
for any majority language page, unless dictionary analysis is available 
*and* the font contains OTL language system tagged variant behaviour 
appropriate for majority use.

Is dictionary analysis the sort of heuristic you are envisioning?


JH
Received on Tuesday, 5 February 2013 19:01:24 GMT

This archive was generated by hypermail 2.3.1 : Tuesday, 26 March 2013 17:21:05 GMT