Re: Possible issue: Accept-language priority based on language order

On 2011-11-24 07:41, "Martin J. Dürst" wrote:
> Hello Julian,
>
> On 2011/11/24 7:41, Julian Reschke wrote:
>> On 2011-11-23 23:29, Amos Jeffries wrote:
>>> ...
>>> I've been keeping an eye on this since implementing language negotiation
>>> in Squid.
>>>
>>> It appears that nearly all agents are sending the language codes sorted
>>> by q value anyway. Whether they send the q value or not it is still
>>> possible to optimize by using the left-most wins assumption.
>>>
>>> If anyone is interested in doing a deeper analysis I have a dataset
>>> available covering the last year on several networks linking the
>>> Accept-Language and User-Agent header pair.
>>> ...
>>
>> Analysis would be good.
>
> I agree with Harald's analysis. I think it's up to people who want to
> claim the contrary to do some footwork.
>
> I would definitely NOT go as far as Dale and say "ignore the q= values,
> they will be in order". That would explicitly be against the current
> spec. But saying that if there are no q values, then the leftmost
> matching should win will definitely bring the spec and reality closer
> together.

That's not clear at all. A conforming implementation might handle things 
with the same q value (or missing values) as a set (unordered) and would 
become non-compliant with this change.

> ...
>> I'm also not too enthusiastic having to consider whether this would be
>> *specific* to Accept-Language, or apply to all Accept headers.
>
> I don't know much about the other headers. There is a strong difference
> between Accept-Language (which can be set by the user through
> Options/Preferences in most browsers) and the others (which are mostly
> just baked into the browser).

Indeed. But they share a similar syntax, and I'm really REALLY unhappy 
to add more special cases when they are not really needed.

Best regards, Julian

Received on Thursday, 24 November 2011 08:44:06 UTC