[XeTeX] polyglossia becoming reality...

[Arthur Reutenauer]

> I forgot to mention that I would also welcome contributions of gloss-* 
> files, especially for Indic and East Asian scripts and languages :) But 
> perhaps we should wait until the interface has stabilized.

  I wrote LuaTeX macros to implement the Japanese kinsoku rules a couple
of months ago; it should be possible to adapt them to XeTeX (with the
help of \XeTeXinterchartoks).  And it could make sense to merge
zhspacing into polyglossia, too.

  Oh, and another thought: this could be an opportunity to clean up the
language naming mess by adopting a completely new scheme: what about
tagging the languages with things like ISO 639 codes?  Of course real
names are prettier and nicer than 2- or 3-letter codes, but it isn't
always easy to be as precise as possible.  There is an RFC (4646,
ftp://ftp.rfc-editor.org/in-notes/rfc4646.txt) which attempts to unify
all the different language variants into a single set of tags, taking in
account spelling reforms, differences in writing systems, regional
variants and all sorts of other things.  For example, it defines tags
for the current spelling reform of German (de-1996) and the old spelling
(de-1901), and we could want to support both (ConTeXt does that,
following requests from German users); if you want to mention the name
of the script you can do that (the relevant standard is ISO 15924), so
RFC 4646 defines zh-Hant and zh-Hans as the tags for Traditional and
Simplified Chinese, respectively (which is in my opinion much more
satisfying than using the country codes CN and TW, like in POSIX
locales).  And of course you can still add the country code for national
variants, if needed (so one has pt-PT and pt-BR for Portugal Portuguese
and Brazilian Portuguese, for example).

	Arthur