I've noticed that some sentence is missing in myGengo. Furthermore, some of these was present under svn.
How does the c5 team generate the language templates?
IMO the best approach would be to create a script that extracts the sentences from php files and from some of the DB table fiields of the initial data (we should parse the two sql files used in the setup process). Here I'd use xgettext to create the pot file.
After that, we should update the po files of each language with msgmerge, and update myGengo.
If you need, I can help with this process...

Ciao!
Michele