-*- outline -*-

Guess-lang automagically guesses what is the language of a buffer.

This works basically by counting occurrences of common words in every
language that is known and comparing respective numbers.  Languages
currently supported are english, french, german, danish, polish,
spanish, esperanto and somewhat russian (based on loosely collected
words from various .po files ... does anybody want to help ?).

guess-lang.el is based on an idea from Pascal Courtois
<Pascal.Courtois@nouvo.com> but evolved a lot.


* How to install it ?

See INSTALL, it is pretty easy.


* How to add support for a new language

- collect enough (>200k) text of your language.  This must be of
  general interest and must use ``standard'' style (i.e. excerpts from
  newspapers).  Mails are fine provided you remove headers,
  signatures and take care that no particular words remain in your
  final texts (how strange ``linux'' is common in swedish ;-)).

- save all of your text in a large file.

- run compute-frequencies.pl (which can be found in guess-lang 1.9
  and above) on it.  Use 'compute-frequencies.pl --help' to now what
  options are available.

  Example: ./compute-frequencies.pl -f my-words

- save output from compute-frequencies.pl inside a file.

- send it to bdrieu@april.org, I will include it in next release of
  guess-lang.

- copy it to ~/.guess-lang/<language>, where language is the name
  of your language.

  Example: /home/benj/.guess-lang/french

- add your language in the guess-lang-languages-to-guess variable.

  Example: (setq guess-lang-languages-to-guess '("francais" "british"))

  Hint: use `M-x customize-group guess-lang' for that

- you can optionnaly define aliases for your language by fiddling with
  the guess-lang-dictionnaries-aliases variable.


* How to add a new coding system

Unless you feel like it, don't bother.  Just complain to me
(bdrieu@april.org).


* Integration with ispell

[To write]
