[OS X TeX] To XeTeX or not to XeTeX (was: gtamacfonts ligatures: PDF searchability)

Robert Spence spence at saar.de
Tue Mar 28 15:23:37 EST 2006

Apologies for having started a thread within a thread by sending a  
mail about font licenses that arose out of my question concerning  
gtamacfont ligatures.  (Hopefully the mailing lists of the future  
will be organized more like CMaps or something, so that the mistakes  
of the technologically challenged will have a chance of going  

On 28 Mar 2006, at 15:16 , Bruno Voisin wrote:
> Le 28 mars 06 à 14:47, William Adams a écrit :
>> The solution here is to use XeTeX, and to improve it and support for
>> it, so there're no reasons not to use it.
> But there are at least two reasons for which some people can't use  
> XeTeX right now as their main TeX platform (apart from the speed  
> issue, which isn't so important IMO):
> - The huge number of LaTeX packages which assume you are either  
> using dvips or pdfTeX, and have no driver file for XeTeX (or, even  
> worse, bypass the driver file mechanism by hard-coding driver- 
> specific instructions within the .sty package file itself). Think  
> of all the users of beamer, for example. And I'm not even speaking  
> of packages, like pstricks, which rely entirely on dvips.

For me it would be more of an opportunity to learn about the one  
route I _haven't_ systematically tried yet, dvipdfm.  I feel  
comfortable enough about switching between the dvips route and the  
pdftex route as far as graphics inclusion is concerned, and only  
decided to go with pdftex in the end because it could do better lines  
around boxes, and time constraints had prevented me from doing too  
many things with pstricks anyway.  Luckily.  But the beamer  
incompatibility is going to be a problem.  When I first saw beamer I  
thought: "This is it!"  Then, while playing with it, I got  
sidetracked by two questions:
1. How did its designer get another Chinese font than Cyberbit.ttf  
into the Chinese example?
2. Why not add something like \AcrobatMenu{Find}{?} to the navigation  
bar I'd designed for the footer of my documents?
Pursuing 1) brought up the issue of Type 1 versus OpenType; pursuing  
2) brought up the issue of just what it was that was being searched  
for (and found, or not found) in the PDF document; and both together  
led to my initial question to the list.  (Question 2 also raised some  
issues with the dvipfm driver, which didn't get all the hyperlinks  
right, but I can postpone that, because the navigation bar is  
superfluous without beamer anyway, as it implies you're in fullscreen  
mode and have cropped the pages accordingly.)
I guess my current feeling is that I could afford to be without  
beamer until no later than a couple of weeks before the next time I  
have to give a conference presentation.

> - The way XeTeX deals with Mac fonts, letting the Mac OS X font  
> mechanism set the baseline. This results in uneven line spacing,  
> especially in maths documents with inline subscripts, superscripts  
> and the like, and makes the output unsuitable for wide  
> distribution. (I know there are workarounds, involving fussing with  
> \linespread IIRC, but they are not perfect and require a fair level  
> of familiarity with plain
> TeX internals.)

TeXShop's PDF display is so good that I noticed the problem with the  
baseline as soon as I made my first attempts at getting XeTeX to  
change fonts midline.  That, and the fact that there were some low- 
level spacing commands in the example sources, started to make me a  
bit worried, as I know (from attempting to learn l2tabuen off by  
heart) why one should leave that kind of work to packages.  I was  
also trying to do some unorthodox things with fancyhdr and geometry  
at the same time, and using scrartcl options instead of the parskip  
package, to get the page layout right for what was going to be a new  
local .cls file by a week ago and is still a mess, so I was already  
sensitized to low-level typesetting problems.

So far I've watched XeTeX demolish a (fancy, badly planned, sloppily  
coded) document template test which I'd already made with pdfetex and  
which theoretically should have compiled in "legacy" mode;  then I  
started from scratch with something simple and added one package at a  
time to check out what incompatibilities there are and what  
workarounds are possible.  Current state of play is: optimistic.   
Provided I don't run out of time, I think I can work out most things  
by myself, without having to plague the XeTeX list with too many  
stupid novice questions.  But I think I would feel very uneasy about  
producing any documents at all in which the leading wasn't right!

Not being a programmer, I couldn't really make much of a real  
contribution to the XeTeX project (other than a financial one,  
perhaps), and as a linguist my debt to SIL is of course already  
incalculable (its founder's wonderful theory of language; all those  
free fonts and other tools; ...).  I could perhaps help with  
documentation --- proofreading, or translating, or something.  And  
file bug reports, of course.  I think XeTeX is definitely worth  
investing some time in.  (Whether my students will think so when  
lectures start in a couple of weeks is another matter.  They don't  
want empty templates or well-designed document classes with no  
documents written in them.)  So I'll take the plunge, as it's quite  
educational to find that I can't scale OS X system fonts to  
MatchLowercase without first getting the _latest_ version of  
fontspec, which then complains I haven't got the right version of  
xkeyval, and so on.  This is what TeXing is all about.  Working with  
Word under Windows always made me feel like an idiot; working with  
TeX on some flavour of UN*X, with or without all the iCute stuff,  
makes me feel like a floundering beginner.  The difference is small,  
but important.

It would be nice, though, if there were little plateaux of  
developmental stability every now and then --- maybe XeTeX version  
1.0 ? --- to give the learning curve of beginners a chance to flatten  
out a bit.  But I guess if it's possible to survive the transition  
from Mac OS 8 to Mac OS X Tiger, or from 7-bit-encoded bitmap fonts  
to 8-bit-encoded Type 1 fonts, then it's possible to survive almost  

So count me in!


--- Robert Spence
Applied Linguistics
Saarland University
"Maybe they wouldn't notice the small caps for highlighting technical  
terms were Hoefler Text not Baskerville if we colored them a discreet  
[rgb]{0,.32,.65}"------------------------- Info --------------------------
Mac-TeX Website: http://www.esm.psu.edu/mac-tex/
          & FAQ: http://latex.yauh.de/faq/
TeX FAQ: http://www.tex.ac.uk/faq
List Archive: http://tug.org/pipermail/macostex-archives/

More information about the MacOSX-TeX mailing list