W3C home > Mailing lists > Public > public-appformats@w3.org > January 2008

Re: Examining the 'no server modification' requirement

From: Mark Nottingham <mnot@yahoo-inc.com>
Date: Tue, 15 Jan 2008 11:27:36 +1100
Cc: Brad Porter <bwporter@yahoo.com>, "WAF WG (public)" <public-appformats@w3.org>
Message-Id: <5E79BE16-2ECC-4D52-9A83-2D4A72667F2D@yahoo-inc.com>
To: Thomas Roessler <tlr@w3.org>

If the 4xx were cacheable, it would be served from cache; if not, it  
would go back. Either way, the right thing happens (as long as all of  
the headers that affect the response are listed in Vary).


On 13/01/2008, at 5:47 AM, Thomas Roessler wrote:

> On 2008-01-11 17:15:03 +1100, Mark Nottingham wrote:
>
>> That's the beauty of the server-side model; it works very well with  
>> caching.
>>
>> E.g., if the request is
>>
>> GET /foo HTTP/1.1
>> Host: www.example.com
>> Referer-Root: http://other.example.org/
>>
>> The response could be
>>
>> HTTP/1.1 200 OK
>> Cache-Control: max-age=3600
>> Vary: Referer-Root
>>
>> ...
>>
>> which tells a cache that it can serve that response to other
>> clients, *as long as* they send the same Referer-Root header. The
>> cache ends up enforcing the server's policy on its behalf,
>> without any new software.
>
> If a 4xx response was seen before, the request would still go back
> to the original server, right?
>
> Thanks,
> -- 
> Thomas Roessler, W3C  <tlr@w3.org>

--
Mark Nottingham       mnot@yahoo-inc.com
Received on Tuesday, 15 January 2008 00:27:58 UTC

This archive was generated by hypermail 2.4.0 : Friday, 17 January 2020 18:56:21 UTC