[XeTeX] combining overline and other unicode features

Jonathan Kew jonathan_kew at sil.org
Sat Jun 25 18:29:50 CEST 2005


On 25 Jun 2005, at 4:21 pm, Bernd wrote:

> Hello,
> I've been playing around with the transliteration of Arabic using  
> Unicode Characters. I found out that combining characters for dots  
> and bars exist but I have no idea how to apply them. They copy fine  
> from other texts but how do I type these myself? Is there a  
> combination I have to enter on my keyboard as you do with accents  
> and graves on vowels? Or do I need to enter the unicode number plus  
> the letter?

It depends on the keyboard layout you're using, and what characters  
that is able to generate. The standard US layout doesn't provide  
keystrokes for these, but others might. Besides the layouts that ship  
with the OS, there are links to some other available layouts,  
including some intended for phonetic transcription, at:

     http://scripts.sil.org/inputresources

If no available layout is suitable for your needs, there are several  
utilities to help you build your own custom layout. One you could try  
is Ukelele:

     http://scripts.sil.org/ukelele

Alternatively, you can enter any Unicode character using the Hex  
Input layout that Apple provides (and which you can enable via the  
International preference panel); using this, you can type a 4-digit  
Unicode value (in hex) while holding down the Option key, to generate  
a single character. So <Option-0-3-0-4> gives the combining macron  
(bar) above a letter, for example.

Or you can encode the characters in your TeX source using the  
^^^^xxxx notation (four ^ characters, followed by four lowercase hex  
digits); so ^^^^0304 also represents the combining macron.

Note that how well these combining characters work depends whether  
the font you're using has full support for placing them properly over  
the base characters; with many fonts, you may not get good  
positioning. One that should support most combinations well is Doulos  
SIL:

     http://scripts.sil.org/doulossilfont

With many fonts, you'll get better results if you use precomposed  
Unicode characters that represent the base letter + diacritic as a  
single code, for the cases where these are available in Unicode.  
That's what the standard keyboard layouts generate when you use "dead- 
keys" to apply accents like acute/grave; they don't use the  
individual combining marks, as relatively few fonts support these  
fully. A tool like Ukelele would allow you to create a keyboard with  
additional dead-keys for the precomposed accented characters you want  
to use.

Hope this is helpful,

JK



More information about the XeTeX mailing list