Reading:

• http://docs.python.org/lib/node738.html
• http://po-utils.progiciels-bpi.ca/showfile.html?name=dist/README

What do I understand?:

The process of i18n (and l8n) involves a couple of steps:

decide a name for the application (e.g. myapp)

1. ‘marking’ string in the source as translated, most of the time using the underscore function

2. Collecting all the string into a translation template - *.pot (pygettext)

3. translating the files - basically copy the .pot into a .po file, and then translating

4. converting the .po into a .mo file (msgfmt.py)

5. copy the .mo file at the proper location locale/de/LC_MESSAGES/myapp.mo
   locale/de/LC_MESSAGES/myapp.mo
   locale/london/LC_MESSAGES/myapp.mo

6. In the source, do a:

   import gettext 
   lang1 = gettext.translation('myapp', './locale',languages=['de']) 
   lang2 = gettext.translation('myapp', './locale',languages=['london','en'])
   lang1.install() 
   print _('foobar') 
   lang2.install() 
   print _('foobar')` 

That should be it