Spell Checking and Dictionaries
One of the goals of the Lingucomponent project is to develop dictionaries and affix files to support spell checking in different languages.
Dictionaries/Affix files Now Available
Please see the Dictionary Download Page for more information.
How to Create Dictionary and Affix File for Another Language
The MySpell spell checker uses a modified version of Ispell's dictionaries and affix files (modified to permit fast parsing, to be case sensitive, etc.) so that "munched" dictionaries can be kept in straight text. The same is true for Hunspell which has replaced MySpell in OpenOffice.org 2.0.2 and later.
So, the simplest way to add spell checking capabilities is to:
1. Find an Ispell dictionary and affix file (both have to be modified to work). See http://fmg-www.cs.ucla.edu/geoff/ispell-dictionaries.html for a good starting point.
2. Contact the author of the wordlist to make sure licenses are compatible. A BSD or LGPL license is compatible. If the license is not compatible you may want to ask the author for permission to use the wordlist and affix file under a special non-GPL license given that OpenOffice.org is LGPL, is free, is open source, etc. Remember to retain the original e-mail of the author granting permission, because it will be needed in a later step.
3. Convert the affix file and wordlist from the Ispell formats to the formats needed by MySpell/Hunspell. Converting an affix file is time consuming (several hours) but no programming skills are required and all that is needed is a simple text editor. The best way to understand how to translate an affix file to MySpell format is by example: Ispell's English Affix File and MySpell's English Affix File and then grab the following README (text file) that explains in more detail what must be done. Also check out the Hunspell homepage. Hunspell supports MySpell syntax but offers several extensions which you might need.
4. Download and test your affix file and dictionary using the Hunspell stand-alone realease. If you want your files to work with Myspell, use a standalone copy of MySpell (320KB) which can be compiled simply on most platforms (should work fine under Unix, Solaris, Linux, etc).
5. Create an issue (you must be logged in) for the lingucomponent project asking the developers to include support for spell checking in your locale (en_US, en_CA, de_DE, pt_PT, etc) and supply the modified wordlist and affix file. You must also include a README.txt file that specifies the author; original ispell author, if any; the specific language and locale supported; and the license you are releasing your dictionary under (GPL, LGPL, BSD, or other).
Once you have reached here: Congrats! You have helped the OpenOffice.org effort - Thank You!
Last Modified: $Date: 2006/04/06 20:05:01 $, $Revision: 1.17 $