How about using Weblate? #1336

[meijer]

Community asked for a decision on potentially using Weblate for automated translation and localisation:
https://weblate.org

Ticket in dev repo: filesender/filesender#1336

[meijer]

https://weblate.org/en-gb/hosting/

"Weblate proudly supports libre projects

If you feel like being supported by Weblate will help you, set up your libre project and get the Libre plan gratis. It has the same limits as the Advanced plan, but only for public projects."

If that wouldn't come through, cost may be acceptable. Business model scales with number of languages, number of projects and source strings but number of languages we support + source strings falls within reasonable limits.

[WebSpider]

Personally, I have no experience with using weblate, let alone with open source projects.

I usually tend to look at bigger open source projects and how they handle translations, and for instance nextcloud uses Transifex, which I really like for the options they have for collaborative translations, and integration into CI/CD which alleviates the work devs need to do to keep up with translations.

If we switch for some reason, I would like to propose we look at a tool that:

• Can easily be used for multiple backend languages, not just php
• Is easy on the eyes for translators, so it becomes easy to propose a new translation
• Can be driven by github actions, so we can include translations into every release without anyone needing to spend a mental CPU cycle on it

If that's weblate, all good. If that's Transifex, all good, if that's poeditor, equally good.

[monkeyiq]

There is currently a fairly simple method to bring in translated terms from poeditor to github
scripts/language/import-all-from-poeditor.sh

I am not personally attached to any particular service.

[monkeyiq]

The big question is what do other services offer that we would like to have but lack with poeditor.

[WebSpider]

As part of the last gettogether, I have investigated poeditor and weblate a bit, and they are very similar. We can use the (new?) GitHub integration in poeditor to significantly better integrate translation updates into the release cycle, which I'll set up a fork to work out.

The main idea is not to have to think about it, as the release CI handles updating languages from poeditor, and the CI running on develop pushes new terms that need to be translated into poeditor.

At this moment, I don't think the eyecandy that weblate offers justifies the effort of moving to a different translation platform, including moving over users, changing CI, etc. But this may change once I have FS3.0 translation CI (is that TI? ;) ) set up in a testing environment.

[WebSpider]

I guess the consensus here it that we see no clear reason to move away from our current translation solution.