This is a list of things that needs to be done to propagate the changes properly.
Write DONE to each issue once resolved.

Wishlist:
- separate language.dat files on per-engine level (a different language.dat for XeTeX than for pdfTeX)
- different syntax for loading multiple patterns? (Serbian)

Short-term TODO list:
- A web page with simple examples how to enable hyphenation
- Documentation about hyphenation setup in TeX Live 2011, MikTeX & W32TeX
- Documentation about conventions (naming guidelines, no-sequences-in-patterns, stable vs. non-stable filenames, ...) of hyph-utf8
- Put licence in the manual (and explain licences of others)
- Hans' examples for comparisons & documentation
- discuss switching to special patterns for pdfTeX
- better solutions for apostrophe



TeX Distributions:
- [DONE] propagate changes into TeX Live
- [DONE] coordinate with Christian Schenk <cs at miktex.org> for changes in MikTeX
- [DONE] coordinate with Akira Kakuto <kakuto at fsci.fuk.kindai.ac.jp> for changes in W32TeX

Pattern creators:
- make a list of pattern authors, notify them about changes,
  explain them that they need to modify UTF-8 file/submit file it as UTF-8 in future
- write DONE for each language once you receive the answer from the author

Languages:
- complete languages.html
- unification of Mongolian
  - encoding
  - pattern versions
    - [DONE] a list of differences received from Dorjgotov Batmunkh <bataak at gmail.com>
    - add that list to repository
    - waiting for answer from Oliver Corff <corff at zedat.fu-berlin.de>
- unification of Spanish (some new version of patterns submitted)
- describe differences between sh-cyrl and sr-cyrl
- [DONE] do something with EO: I did the conversion for "plain" patterns, but left the original version intact
  we could leave a separate version in "source" and use the "plain version" in pattern repository
- Russian etc. that are still outsourcing
  - solve encoding mess
- versioning: German etc.
- Norwegian
- Belarussian:
  - Contact the authors of hyphenation patterns on OOo:
    http://extensions.services.openoffice.org/en/project/dict-be-classic
    http://extensions.services.openoffice.org/en/project/dict-be-official
  - Contact the author of support files for LaTeX (Aleksey Novodvorsky):
    http://alt.linux.kiev.ua/srpm/tetex/patches/8
    http://tldp.org/HOWTO/pdf/Belarusian-HOWTO.pdf
  - Convert patterns into UTF-8
  - Babel and Polyglossia support
  - Which version has to be used? Classic or official?
  - Do something about encoding in 8-bit engines (possibly together with Vladimir Volovich)

CTAN:
- move old patterns to obsolete
- [DONE] upload new patterns
- make an easy-to-browse structure
- ask CTAN people to be careful if someone uploads new patterns
- write auto-notificator for changes in CTAN
- try to update the information in TeX Catalogue
  - http://www.tex.ac.uk/tex-archive/help/Catalogue/hier.html
  - some entries say that patterns are obsolete, but still claim that the package is included in TL/MikTeX
    (http://www.tex.ac.uk/tex-archive/help/Catalogue/entries/hrhyph.html)
  - [DONE] change script 4TL to generate shortdesc for hyph-xxx.tlpsrc
  - once the shortdesc is written, also make sure to put longdesc or more extensive discription for TeX catalogue
  - should we release a CTAN-only "browsable" version of hyphenation patterns:
    - one folder per language
    - symbolic links to all the files for that language
    - a nice README for every language
      - lists the author, copyright, ...
      - says that the package is part of hyph-utf8 and properly redirect
    - catalogue could have a reference to that folder and contain all the data
    - discuss the idea with CTAN before doing extensive changes

OpenOffice:
- write auto-notificator for changes in OOo (more important)
- check every single language on http://extensions.services.openoffice.org/en/dictionaries
- contact the authors & join the efforts
- [DONE] import indic scripts

Impromevents:
- prepare a better version of notes
- auto-conversion into four different files:
  - licence
  - list of used letters for each language (lowercase and uppercase equivalent); difference between letters & others
  - [DONE] list of patterns
  - [DONE] list of exceptions
- that list should be auto-generated from TeX sources and used for [complete a list]:
  - hyphenator
  - ...
- LuaTeX patterns
- [almost DONE] Describe and try to solve the apostrophe problem (users should be able
  to input both U+0027 and U+2019)
  See http://tug.org/pipermail/xetex/2010-October/018914.html ff. for
  one use case.

Documentation:
- authors, history, changes
- how to use patterns in plain TeX
- how to use patterns in LaTeX with polyglossia (or babel?)
- how to use patterns in ConTeXt
- how to add a new language (once we quit the project)

Testing:
- collect lists of words for the languages where that is possible (no copyright problems)
- write a luatex tester for a collection of words for each language

Licences:
- ... complete the list with tasks that need to be done
- we want two licences: one for content and one for packaging
- write a simple licence for packaging:
  - we don't care what others (outside of TeX world) do; just mention the credit to hyph-utf8
    and complete credit to pattern authors
  - respect the licence of pattern authors
  - within TeX world:
    - keep the functionality; implementation may change, but not the end result (unless improved)
    - contact the list first when changes are needed
    - if team disappears, one should be free to continue the work (mention how that should be done)
    - some general guidelines:
      - no TeX commands (apart from \hyphenation and \patterns)
      - everything in Unicode
      - hyph-xx.tex will keep containing the patterns
      - loadhyph-xx.tex will keep loading the patterns, but may change considerably (ugly tricks should go there)
      - ... (complete the list)
    - language-specific:
      - author of existing patterns has full right to submit changes
      - when a new author comes and the old one doesn't respond ... what to do?
      - when two version of patterns appear, try to convince the authors to combine the patterns
        ... write some general guidelines
- collect a few ideas for valid licences (maybe a different licence f)
- write the authors if they agree with that licence (give them a few options to choose from)
- write some really nice, easily parsable form and apply to all the patterns
- modify generating scripts for:
  - German
  - gl
  - eu, tk, tr

Example of a licence:

This work consists of two parts:
a) patterns for different languages
b) support files (tex files, generating scripts, documentation, ...)
with two different licences for each:
1) for within the TeX world
2) and outside the TeX world

A short version of licence:
b2) We don't care what you do, though it would be nice of you to mention the source of patterns (hyph-utf8 residing at http://tug.org/tex-hyphen) at some visible place.
b1)

a1)

Collaboration:
- make a list of people that we would want to collaborate with
- some notification system for those using our patterns as source for whenever something in repository changes
- write done for everyone we have collaborated with

TeX users:
- [on the way] write DOC about pattern conversion
- [DONE] write an article for TUG boat and/or other TeX magazines
- [DONE] maybe present the main idea at some TUG conference
- notify mailing lists
- CTAN people should take care for new patterns to be submitted in the proper form

Web page:
- split the page into several subpages:
  - short intro & news
  - languages - list all available language and add proper links
  - collaboration and links
  - description of algorithm (with Mathias' slides)
- nicer design
- apply Mathias' Hyphenator.js
- add a test page to do hyphenation on-the-fly using Mathias' code
  and creating some really nice output
  - a dropdown box chooses the language
  - text input field
  - button for submission (maybe we don't even need that)
  - sample output

Software:
- possibly rewrite patgen to handle UTF-8
  - [DONE] try to convince Mathias Nater to do that; Deadline: BachoTeX 2011 :)
  - try to convince him to finish the project

OpenOffice:
- http://extensions.services.openoffice.org/en/project/RomanianDictionaryPackCedillaVersion
- (kurdish) http://extensions.services.openoffice.org/en/project/kitandin

- http://www.mail-archive.com/dev@lingucomponent.openoffice.org/msg01104.html



Hello Mojca,


On 16/03/2010 10:21, Mojca Miklavec wrote:
Hello François,

Not so long ago the author of  Mongolian patterns has asked us for
"help about how to use his patterns in TeX Live 2008/09". (They are
called "mongolian2a" in language.dat since there are two version of
Mongolian patterns in TeX Live - both different patterns and different
encoding. I'm not sure that I know how to enable those patterns in
LaTeX in some nice way.)

First question: does polyglossia work ok with 8-bit engines (pdftex)
as a replacement for babel?
  

No. It assumes Unicode encoding and is currently for XeLaTeX only.
Together with Elie Roux I plan to add support for LuaLaTeX in a later version (not very soon however).

Second question: would you be willing to write a short introduction
(in tex or maybe even better in html forhttp://tug.org/tex-hyphen/)

about how to use some particular patterns in polyglossia (maybe we
should start promoting polyglossia via our webpage, in particual if
the answer to my first question was positive).
  

Sure.

The short answer is:

Since Babel's hyphen.cfg is built in the xelatex format, hyphenation patterns can be used without even loading polyglossia (or Babel for that matter). At the low-level this simply corresponds to defining
   \language=\l@<langname>
(where langname is the string identifying a particular hyphenation file in language.dat). The user command for this is
   \hyphenrules{langname}
or
   \begin{hyphenrules}{langname} ... \end{hyphenrules}.

The above works with any flavour of LaTeX, and I think it is also available in Plain TeX with
   \hyphenrules{langname}
   ...
   \endhyphenrules



In particular, we need to provide examples of usage in {plain TeX,
LaTeX, XeTeX, XeLaTeX, ConTeXt, Lua(La)TeX} for different encodings. I
can write something about ConTeXt, but I need help for different
flavours of LaTeX.
  

Under "plain" pdftex, xetex, and luatex, one can use the macro \uselanguage{langname} for any language mentioned in language.def. It's that simple!

Under polyglossia this is all done automatically, as in Babel, within the individual language definition files, depending of the options associated with each language. Example:

\usepackage{polyglossia}
\setmainlanguage{asturian}
\setotherlanguage[variant=usmax]{english}% American English with extended hyphenation patterns
\setotherlanguage[spelling=new,latesthyphen=true]{german}% German with experimental patterns "ngerman-x-latest"
\setotherlanguages{spanish,catalan,french}
\begin{document}
Long Asturian text ... (Hyphenation for Asturian is not available, but polyglossia automatically falls back on Catalan for now, which seems to be a reasonable choice.)

\begin{german}
Deutscher Text ... (with the hyphenation patterns selected above: "ngerman-x-latest")
\end{german}

\begin[script=fraktur,spelling=old]{german}
Deutſcher Text ... (set in Fraktur, with traditional hyphenation).
\end{german}

etc.

Feel free to adapt and use the above to your liking!

Regards,
François