Re: p6: Returning the freshest response

On 30/03/2013, at 3:39 AM, Martin Thomson <martin.thomson@gmail.com> wrote:

> On 29 March 2013 09:02, Ken Murchison <murch@andrew.cmu.edu> wrote:
>> The current language in
>> https://svn.tools.ietf.org/svn/wg/httpbis/draft-ietf-httpbis/latest/p2-semantics.html#proactive.negotiation
>> (MAY) and
>> https://svn.tools.ietf.org/svn/wg/httpbis/draft-ietf-httpbis/latest/p2-semantics.html#header.vary
>> (SHOULD)
>> don't seem to make it a requirement.  However, the SHOULD certainly suggests
>> best practice.
> 
> I think that this might trigger a review comment:
> 
> The MAY is spurious, no 2119 language is needed here: the text need
> only highlight that it is possible (as opposed to permissible) for a
> Vary header to be present.

There seem to be a lot of MAYs in p2; some of them are mostly harmless, but this one seems to be actively misleading. Roy, Julian?


> The SHOULD is qualified sufficiently that I believe that a MUST is
> more appropriate.

The requirement is of the form "SHOULD... unless...", with the clause:

> , unless the variance cannot be crossed or the origin server has been deliberately configured to prevent cache transparency.

so I think this one is OK.


Cheers,


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

Received on Friday, 29 March 2013 22:12:09 UTC