[XeTeX] The future of XeTeX

Ulrike Fischer news3 at nililand.de
Wed Aug 1 12:13:44 CEST 2012

Am Wed, 1 Aug 2012 09:32:16 +0200 schrieb Keith J. Schultz:

> LuaTeX has a very small developer base and their goal is very
> high. a long needed rewrite of TeX. That is a complex task.
> From the simple user side. LuaTeX is about as easy as it gets. For
> most purpose I can teach you all you need to know how to use Lua
> for TeX in 2 hours!

The problem are the people inbetween: The people who should develop
the code needed on top of the binary. As you could see in this
discussion the core problem currently is the handling of (open type)
fonts. And while the fontloader lua code in context (the source of
luaotfload) is quite advanced, it is undocumentated, has no sensible
api, and can change all the time in unexpected ways. Nobody outside
the context team can actually work on it and e.g add support for
scripts or correct bugs. 

(As an aside I think that one should not only put pressure on
xetex/luatex/open type engines to support all sorts of open type
features and scripts but also on some scripts to adapt a bit to the
computer age.) 

> It's price for unicode support and using fontspec. But,
> those ancient packages using encodings should be a thing of the
> past, IMHO.

Well in case of chess fonts they are not "a thing of the past". Not
because of some deficiency of luatex or xetex but because most
glyphs used e.g. by chessboards are not in unicode. You need some
local encoding to access them in a standarized way, and this means
you need the ability to reencode fonts. This should be possible with
luatex (and is in my eyes one of the advantage compared to xetex)
but can't be used due to the unclear state of the font loader.

Ulrike Fischer 

More information about the XeTeX mailing list