Re: Issue 113, was: Proposed resolution for Issue 13 (language tags)

On mån, 2008-08-04 at 08:17 +0200, Julian Reschke wrote:

> But apparently servers choose a default anyway. So the defaulting may be 
> different for different Accept-* headers.

Yes, defaulting is different for different Accept-* headers. Implemeters
are free to select how to handle the case when there is no acceptable
variant. Some may do very elaborate guessing, some just use a default,
some return 406...

For most it's very natural and common sense to return the servers defaul
language if none of the available languages is indicated as acceptable
by Accept-Language. The result will still render fine, even if the user
probably won't understand much.  But it's not the same for most other
Accept-* headers as unacceptable type or encoding most likely will
result in a popup for download in the browser...

But there is no difference in the spec. Implementers are free to use
common sense on when to use a default and when not, and clients should
indicate their preference and not assume all servers have the same
default behavior.

> So again, what problem do we want to solve here?

No idea.. only jumped in to explain what the specs says which aparently
wasn't understood and have not followed the discussion before. (language
is not really in my interest, but cache and negotiation model is)


Received on Monday, 4 August 2008 21:08:44 UTC