Fluent with XLIFF


This initiative looks promising, especially for one who came from Russian culture and have seen so many English-like Russian content in UI :slight_smile:

At the same time, I didn’t get how Fluent files are integrated with traditional localization files (XLIFFs). It seems like Fluent has a completely different paradigm which requires some additional steps to perceive it from the “segment” view.

Do you have any docs about connecting Fluent to a traditional language industry tools?

Hi Sasha,

thanks for your interest.

When we made the decision to go for the FTL file format, we considered using standardized containers for Fluent strings. We decided to go for a dedicated file format instead, because the strings themselves are more involved than the container. And most containers are tougher on the eye than FTL turned out to be.

I’m not aware of anyone that invested in XLIFF to transport Fluent strings so far. The use of that would probably depend on the workflow people use XLIFF for. For a translation tool, the value would be small, as you’d end up with no support for Fluent features, AFAICT. This is similar challenge when using XLIFF for MessageFormat strings. I just quickly googled a bit around that right now, and it seems that folks just package the full complexity of a MF strings into a translatable unit.

As for how to hook up Fluent to translation tools, Matjaz blogged about how we have done it in our in-house tool Pontoon on https://blog.mozilla.org/l10n/2019/04/11/implementing-fluent-in-a-localization-tool/.

We’re also talking to external tool developers right now, and one has already made great progress in their implementation.


1 Like