Re: p6-caching: commentary from -05 to -06

On 25/03/2009, at 8:01 PM, Yves Lafon wrote:

> On Thu, 12 Mar 2009, Mark Nottingham wrote:
>
>> 3.2 Warnings
>> Most Warning-related text moved to Warning header definition
>> Downgraded most requirements for sending and displaying Warning to  
>> SHOULD
>
> It is good to keep some of them as MUST,
> 214 Transformation Applied is still a MUST which is a good thing.
> However 110 Response is stale should be back to MUST instead of  
> SHOULD, otherwise the client doesn't have a way to differentiate the  
> response. (but I know that IRL it's almost never done).

If the response is stale, shouldn't that be detectable regardless by  
calculating its freshness lifetime and age independently?

Cheers,

--
Mark Nottingham     http://www.mnot.net/