explainers-by-googlers / reduce-accept-language

This repository hosts explainer for reducing passive fingerprinting in the Accept-Language header.
Creative Commons Attribution 4.0 International
16 stars 0 forks source link

Content-Language in the wild #11

Open foolip opened 1 month ago

foolip commented 1 month ago

https://github.com/explainers-by-googlers/reduce-accept-language?tab=readme-ov-file#no-content-language-in-the-response-header says:

No Content-Language in the response header: This means the content is intended for all language audiences.

Is this really the case? What proportion of pages send a Content-Language response header, and does its absence imply anything about the language?

As part of a new negotiation approach with the Avail-Lanuages header it doesn't matter so much what existing content does, but as part of an analysis of site compat, I'd be surprised if "no `Content-Language → no problem" is a safe assumption.

Tanych commented 4 weeks ago

We don't assume "no `Content-Language → no problem". For now, most sites actually don't sent the content-language header to indicate the response language, and only small portion of sites actually care about the accept-language header.
If sites would like serves users the best representation based on the accept-language headers, they need to indicate what language they supports, and the browser can pick the best one during the language negotiation. This means sites need to take some action to get users' non-first language preference. If sites only care about users' first language preference, its no-op.