KapiX
commented
9 months ago I agree it would be cool. As you mentioned, UI needs to be changed first. Mockups welcome :)
I would see it as an additional type of view (configured per project), since the existing one works well for apps.
Concerns:
- performance - I did not perform stress tests
- contributor trail - Polyglot only stores contributors for projects in a way that doesn't allow to assign each contributor to the translation they contributed. Only last translation version stores who contributed it. However, even if there are no current translations by someone, if they ever added/modified one, they will still be listed. This was okay so far, but maybe for Haiku there are different requirements. Those need to be discussed. (Some systems use Git, but I'm not sure this is the best option.)
- perhaps some kind of info whether someone is currently editing the translation will have to be implemented.
Work to do before migration starts:
- APIs for mass uploads. (AFAIK Pootle gets one ZIP file with all texts.)
- Job queues - files need to be processed asynchronously when there are a lot of them uploaded at
- Suggestions
On IRC, during a rant/brainstorming session about the fate of Polyglot and other translation problems, @waddlesplash had this idea, why don't we use Polyglot to host Haiku core translations? Surely it would require some plumbing and front-faceting, but it would really be cool!
Interface requirements are a different thing however, a big UI revamp is needed...