Tails

#2 Updated by intrigeri 2019-09-12 09:44:40

Here are the problems I understand this proposal would solve:

• Resources used by extra languages, such as:
  • more work to do for cron.sh
  • bigger database
  • slower Weblate
• More choices in the web UI makes it more difficult to use for active translators.
• Having inactive users with Reviewer credentials violates the principle of least privilege.

Is there any other problem this proposal is would solve?

And reciprocally, the drawbacks of removing non-maintained languages from Weblate are:

• Disabling a language takes work, including:
  • Coordinate removal in tails.git vs. Weblate (we need to delete tons of PO files) vs. Puppet.
  • I’m a bit worried about all the complexity involved in disabling a language. Figuring out how to do it right seems non-trivial and error prone.
  • Remove Reviewer credentials from users who used to do reviews for that language.
• Messy Git history if we ever re-add the language
• If new translators for that language come back later:
  • They’ll need to figure out that they can ask us to re-enable it.
  • Salvaging past work, so they don’t have to start from scratch, may be non-trivial.

I’m not sure it’s worth the effort and risks at this point. In the current state of my thinking, with the info I have at hand, I would say: drop target version, drop “Deliverable for”, and come back to this whenever we’ll really feel the problems caused by having inactive languages on Weblate.

Thoughts?

#5 Updated by Anonymous 2019-09-13 09:10:05

Fully agreed.