Google is causing incalculable damage to the global web with its search indexer not respecting Accept-Language and Content-Language and encouraging webmasters to detect language preferences by IP geolocation instead of respecting the user agent's explicitly provided preference.
Example: just because I am accessing a Google site from an IP geolocated in Japan does not mean I want content localized in Japanese. Especially if my browser is already telling the web server I want it in English. Which it is.
This is particularly silly because Chrome itself even handles the Google Translate integration by observing a combination of Content-Language and the lang attribute of the HTML tag. So some parts of Google seem to recognize this standard, while the search indexer uses fucking content heuristics and separate IPs to detect content language, and totally ignores the standard headers and html features.
