Thought this might be of interest since you have Aspell dictionaries in
your distribution. Sorry for posting it to the list, there did not seam to
be any single maintainer of the Aspell packages.
--
Kevin Atkinson
kevina at users sourceforge net
http://www.ibiblio.org/kevina/
-- Forwarded message --
Date: Fri, 15 Jun 2001 06:08:01 -0400 (EDT)
From: Kevin Atkinson [EMAIL PROTECTED]
To: [EMAIL PROTECTED]
Subject: Official Non-English Word Lists Packages Now Available
[Please redistribute this announcement as you see fit]
In an effort to make installing word lists for non-English languages
in Aspell straightforward I have decided to release foreign language
dictionaries in a standard format.
A preliminary version of my efforts are currently available at the
Aspell home page (http://aspell.sourceforge.net). There you will find
support for the following languages: Breton (br), Catalan (ca), Czech
(cs), Danish (da), Dutch (nl), Esperanto (eo), Faroese (fo), French
(fo), French (fr), German (de), Italian (it), Norwegian (no), Polish
(pl), Russian (ru), Spanish (es), Swedish (sv).
Please check them out and let me know what you think, but please keep
in mind that these are preliminary and subject to change. These
packages contain everything needed to add support for a given language
to Aspell. These packages will also install the necessary files so that
the Word List will be correctly recognized by Pspell -- something that
is often not handled correctly by Word List author's Aspell packages.
Support for variants in languages (such as American, British,
Canadian, Swiss German, etc...) is not yet available as I am currently
not sure the best way to do this. However, support *will* be available
by the next version of Pspell and Aspell as I am planning on also
packaging the English dictionaries this way and distribute it
separately.
Authors of the word lists used to create these packages are encouraged
to check them out to make sure I did things correctly. In the near
future I will allow you (the word list author) to maintain the
packages your self, but for right now I want to maintain tight control
over them as the format is not quite finalized yet.
Also, I am especially interested in feedback from Package Maintainers
(RPM, Debian, etc.) as the Makefile is rather primitive and probably
does not install things correctly nor support the options needed by
maintainers to make package simple. Suggestions are more than
welcome, but patches to the proc script (the script which does all the
real work) are even more welcome. I am also strongly interested in what
you think about the layout of the dictionary files and language names.
Technical Notes for Word List authors
and Package Maintainers
In order to make things as straight forward, portable, and uniform, I
have decided to enforce the following rules.
1) All language names are now the two letter ISO code (en, da) etc.
2) The actual dictionary files must all start with the two letter
code and end in .rws and may only contain ASCII characters.
3) Alias are created using Aspell's multi files and not symbolic links.
These rules are very different from the current way dictionaries are
handled but I fell they will make life easier for everyone.
The reason for the first rule is because in the past the Aspell
language names were a mixture of the name spelled out in English and
the name spelled out in the native language and in some cases involved
non-ASCII characters which was just asking for trouble on non-Unix
like platforms and probably some older Unix ones. Some people want as
far as doing it both ways by symbolically linking one language data
file to the other. This amazingly worked but it is a complete abuse
of how languages names and data files are meant to be used. Finally
others, thought that language variants (American, Swiss German, etc.)
should be considered separate languages and either attempt to specify
them as a language at the command line, for example trying aspell
--lang=canadian ... or creating separate data files for them. All of
this did no good but to confuse people so I wanted to formalize this
and was originally planning on using the language name spelled out in
ASCII characters but released that in many cases I didn't know what
this should be so I decided to go with the universal known language
codes.
The second rule is there so that is is clear which words lists belong
to which languages. I require them to be all ASCII characters for
maximum portability. However, the end user is not expected to use
these words lists directly. Instead they are expected to use one of
the aliases created via the .multi file. These alias can be anything
what so every and may included non ASCII characters. Symbolic links
are not used as there are Unix specific and not supported by Win32.
Non-ASCII characters are okay for aliases as they can simply not be
installed on platforms which
