[XeTeX] Adressing the long s

[Bernhard Fisseni]

Scripsit Jonathan Kew diē 10.8.2008 13:19:
> The best approach use the character U+017F 'ſ' directly in your input  
> text (or if that's not convenient, use a TeX macro that expands to  
> it). That is the Unicode-compliant way to represent the data. Trying  
> to automatically generate 'ſ' where the input text contains 's', using  
> special font features, may seem like a handy trick but will lead to  
> more trouble than it's worth.

Indeed; for German (as in Albert's examples[1]), if you want to follow 
the ‘classical’ rules (short "s" at the end of a syllable, long "s" 
elsewhere) it would also mean that the font (or the font renderer or…) 
would have to do some morphological parsing („Au(-)ſtralien“, unlike 
„Aus(-)tragung“ does not contain a prefix „aus-“[2]) and some kind of 
ambiguity resolution (in the well-known – if rare – 
Wachs-tube/Wach-ſtube cases etc.), or you'd have to use many invisible 
hyphens etc.

For longer (German) texts, it might be interesting to write something 
like rmligs[3] for generating good guesses for long and short "s"s which 
can then be post-edited?  (I might be tempted to get involved if you 
approach this, Albert.)

Bernhard


[1]  I'm not sure of the rules for English, but even if they were the 
same, I suppose that there are fewer exceptions due to the difference in 
morphological structure.

[2] Now that you may separate "s" and "t" in the ‘new’ spelling, this 
particular case may have changed; others, like the "Wachstube" remain. 
And, of course, if one uses the long "s", it will mostly be in cases 
where one follows some original use.

[3] http://www.ctan.org/tex-archive/support/rmligs/