[XeTeX] Long dash "--" in Latex

[Bruno Voisin]

Le 2 juil. 04, à 11:55, Musa Furber a écrit :

> On 2 Jul 2004, at 10:30, Bruno Voisin wrote:
>
>> Le 2 juil. 04, à 03:56, Musa Furber a écrit :
>>
>>> It seems beneficial to provide some sort of compatibility for some 
>>> of the more common LaTeX ligatures, such as `, ``, ', '', -, --, and 
>>> ---.
>>
>> But then you would also have to provide compatibility for accents 
>> like \'e = é, and quotes ` = ‘, ' = ’, etc., and \S = §, and probably 
>> many more like these, and it would end up like reinventing the wheel, 
>> isn't it?
>
> Given my lack of knowledge and experience, I realize that my comments 
> must be severely discounted.

Sorry, I did not mean to sound patronizing, harsh or even contemptuous, 
I was just writing things very quickly being (as always) in a rush.

> Anyway:
>
> The ligatures I mentioned almost always have glyphs defined in the 
> fonts XeTeX will be using.
>
> I wrote "more common" to remove the stranger things, even though this 
> removes things like the strange glyphs I use for transliterating 
> Arabic.

I was just mentioning the letters or diacritics that I have found 
recently to require Unicode input instead of the TeX classical input 
using ligatures and commands. For example:

-- --- \'e ` ' will be displayed using the identical characters from 
the selected Mac OS X font, instead of being replaced by the 
corresponding TeX output characters – — é ‘ ’ in this font. Getting 
these characters in the OS X font requires their direct input as – — é 
‘ ’.

\S will be displayed using the character § from a CM font. Getting this 
character in the selected OS X font requires its direct input as §.

> Perhaps it is a matter of just hooking the spokes up to the wheel, or 
> at least the more obvious spokes.

Yes, provided there are not too many of them. A problem can be that in 
every different language (every different Mac OS X localization) people 
will have a different notion of what are the most obvious spokes (i.e. 
the most obvious diacritics needing compatibility with TeX's standard 
input for them). And this could end up, in some way or another, like 
reimplementing Omega.

Bruno Voisin