[XeTeX] "new-babel", was: Ancient Greek hyphenation

Jonathan Kew jonathan_kew at sil.org
Mon Apr 23 21:39:30 CEST 2007

On 23 Apr 2007, at 8:01 pm, Will Robertson wrote:

> On 24/04/2007, at 4:23 , Arthur Reutenauer wrote:
>>> (3) I agree with you but (4) definitely not -- you *want* to switch
>>> fonts when you switch languages in order to activate any
>>> language-specific OpenType features (although you often won't when
>>> switching between European languages). Don't forget that you need
>>> separate \font declarations to do this.
>>   We need it in the current state of affairs, because as Jonathan
>> pointed out a couple of weeks ago, it's not possible to tell XeTeX
>> “switch to language XXX, using the same font”,

Right - for now, at least. On-the-fly feature changes are a long- 
standing request, but would require some reworking of the internals.  
I don't have any current plans to add this, but it's something that  
might come one day.

>> but is there any other
>> reason why we would like to do it?
> You're unlikely to find a single font that contains high quality
> Japanese + Roman + Arabic letters.

True, but you do find fonts that support several different languages  
using the same script, but with slightly different rendering  
conventions; e.g. Turkish vs Vietnamese vs English (all Latin  
script), or Persian vs Urdu vs Sindhi (all Arabic script). In cases  
like this, it could be helpful to change the language setting without  
having to re-specify the font.

There's a difference, I think, between switching language (which  
affects hyphenation, and possibly some low-level details of  
rendering), and switching script (which happens at some language  
changes, but not all). Maybe this needs to be more explicitly part of  
the model.


More information about the XeTeX mailing list