Is it safe to expose ids?

[JohnHeitmann]

I got vaguely worried seeing the error fallback of outputting message ids as formatted messages in error cases. I don't think developers will be thinking about their ids being customer visible when they are creating them. I admit I'm hard-pressed to come up with a great example of how this might cause harm, but I wanted to make sure the discussion happened just in case.

Mild worries:

• Internal code names in ids
• ids that reflect unreleased secret features, e.g. rebranded-title = ...'
• Ids that seem insensitive without internal/cultural context (or that are just plain insensitive), e.g. dickbar-title = ...

Maybe the Fluent spec itself should call out id-exposing error handling? I'll try to clarify it in the fluent-rs docs.

[zbraniecki]

I don't think developers will be thinking about their ids being customer visible when they are creating them.

This is one of the cornerstones of Fluent design model. Quoting from https://github.com/projectfluent/fluent/wiki/Error-Handling

In the complete breakage scenario where no value can be resolved out of the Message or when no fallback language contains a missing message, Fluent vision is to make bindings return an ID of the message in place of the value, which is a significant improvement for the user over an empty string most translation systems return.

Hope that helps!

[JohnHeitmann]

Yup, that makes sense. It might make sense to also note that behavior briefly in the main reference.