Select and implement a continuous translation service

[tfmorris]

Describe the problem that you'd like solved

There are a number of web-based crowdsourced translation services available which offer free translation to open source projects. We should evaluate the various alternatives, select one and set up OpenLibrary as a project on the platform to receive ongoing translations as the UI is changed.

Possibilities

Some possible candidates include:

• Weblate
• Translatewiki - https://translatewiki.net/wiki/Translating:New_project
• Zanata
• Crowdin
• [Others?]

Evaluation Criteria

• Free to open source projects (mandatory)
• Supports Gettext .po format (mandatory)
• no lock-in
• Open source

Stakeholders

@mekarpeles

[BrittanyBunk]

I would say Google Translate is open source and to have people contribute there, but it seems that you want a company with translators to translate specific writing.

[tfmorris]

I would say Google Translate is open source and to have people contribute there, but it seems that you want a company with translators to translate specific writing.

Google Translate is a commercial Machine Translation (MT) service and most definitely not open source. We want a service powered by native speakers, not a MT service (although some of the human powered platforms use MT assist or first pass).

[BrittanyBunk]

@tfmorris I edit it all the time and there's a section where it feeds a person a line and they translate it. So it's not only by machine, but crowdsourced too. That's fine if it's not open source, but https://opennmt.net/ is the open source version of Google Translate (it's machine based though). Maybe there's a way to get volunteers in various languages too, idk.

[ishank-dev]

@tfmorris, Have you tried Yandex Translation APIs?
They are free to use if used through rapidAPI platform. But again if you are looking for human translation services then I don't think such services would be freely available.

[milotype]

If you're still considering using a translation platform – which I also strongly suggest! – I agree with the suggestions of @tfmorris. Weblate is also my first choice – and it is already used for Internet Archive – but somebody would need to check if it is possible to use it for the project openlibrary. And of course it would probably involve some programming issues …

P.S. I'm not a programmer. I'm a translator.