i have an app i want to translate. Initially before pontoon, what i did was use https://asrt.gluege.boerde.de to generate a bunch of localization resources in different languages. but of course its all machine and not very good. so now i want to use pontoon. what should my strategy be? should i only import the english locale and have people translate from scratch without any of the machine generated translations? Or import all the locales but then all translations would be already prefilled? i like the second strategy better but i am worried that since all terms are translated, everyone will start looking at translations from top down and i think the sort order will be the same for all. so as people onboard, they will all be looking over the same translations depending on what the sort order is? Is there anything i can configure in pontoon so the sort order could be oldest terms first maybe? That way if a term gets updated by a user, it could get sent to the bottom of the view?
Hi @crumplebow3!
For your specific situation, it might be interesting to mark all existing translations as “Unreviewed suggestions”. After you import them into Pontoon, you’d need to run a DB query to set the approved flag to False for all Translation model instances.
You might also want to delete all translation from the translation memory.
As for the sorting, I don’t think we sort files on dashboards or strings withing the translation workbench based on their status. Did you have some other sort order in mind?
Thanks, I tried changing to unreviewed and now when I try to approve a translation in UI I get "django.db.utils.IntegrityError: new row for relation “base_translatedresource” violates check constraint “base_translatedresource_unreviewed_strings_29c744ec_check”
Think I fixed by also removing all translation memories.
EDIT: When I ran a git sync it changed all to approved again 
Anything I can do to not do that?