Hacker Newsnew | past | comments | ask | show | jobs | submitlogin

Accept-Language is clearly the correct solution. It's easy to set in browsers (at least it is in Opera, if it's not in some other browser then I contend that's a UI bug), and OEMs should set it to a sane default for the region in which the browser is distributed.

The indexability is an interesting argument, but my response is that there should be a standard mechanism to query, via HTTP OPTIONS, in which languages a resource is available. By doing so, a user agent or search engine can easily index all versions of a resource.

Whether or not such a standard exists I do not know, but Google certainly has the clout to standardize such a mechanism (viz. sitemaps and #!).



Unfortunately Accept-Language simply doesn't work - if you talk to engineers at Google they'll tell you that they've done the research and an enormous number of browsers have the incorrect setting and hence send an inappropriate header.


How often is an accept-language that isn't set to English wrong?


What would you propose when, as translations have to be done manually, the available content for different pages is out of sync?

For example, maybe that latest product announcement hasn't been translated to French yet. If I send a product list URL to my french-speaking buddy, should the product disappear from the product list?




Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: