[XeTeX] long "s"

Jonathan Kew jonathan_kew at sil.org
Fri Jun 25 22:40:08 CEST 2004


On 25 Jun 2004, at 3:13 pm, Somadevah at aol.com wrote:

> Dear listmembers,
>
>  I have tried but cannot get the right combination of settings to 
> ensure that the long "s" does not appear at the end of a word (as it 
> should not) in Adobe opentype fonts. Can this not be done in XeTeX?
>

Not AFAIK, because Adobe has not implemented the 'hist' feature as a 
contextual lookup, the way Apple has with the "archaic non-final forms" 
or whatever it's called in Hoefler and others.

So to get the long 's' non-final, but the regular form in final 
position, it's up to the application or the user to evaluate which 's' 
occurrences are final, and to apply different feature tags accordingly. 
So in XeTeX terms, you'd have to define two separate fonts, one with 
the 'hist' feature and one without, and write sufficiently cunning 
macros to switch between them appropriately. (And then you'd lose 
kerning at the font changes, I expect, and other issues.....)

All of which is to say that it's really not somewhere you want to go. 
Personally, I think Adobe should have made 'hist' a contextual feature, 
then it would actually be useful. As it is, you can only really make 
use of the feature by manually applying it in individual cases.

Sorry about that!

(Incidentally, this does suggest a possible XeTeX enhancement: a new 
command that can be used to add (or remove) feature tags for a range of 
text, adding to what was given in the \font specification. I'll need to 
think about how feasible that may be; it rather cuts across the 
existing design philosophy whereby a XeTeX "\font" is an instantiation 
of a specific face/size/color/feature-set, and thus provides all the 
input necessary to determine how to render a particular string of text. 
In the long "s" case, though, it would still be up to the user (or some 
clever macros) to figure out where to add the appropriate feature 
override. So still not very helpful of Adobe to do it this way....)

Jonathan



More information about the XeTeX mailing list