W3C home > Mailing lists > Public > ietf-http-wg@w3.org > April to June 2010

Re: Issue 39: proposed example for varying the etag based on conneg

From: Henrik Nordström <henrik@henriknordstrom.net>
Date: Tue, 04 May 2010 01:21:39 +0200
To: nathan@webr3.org
Cc: Jamie Lokier <jamie@shareable.org>, Julian Reschke <julian.reschke@gmx.de>, Yves Lafon <ylafon@w3.org>, HTTP Working Group <ietf-http-wg@w3.org>
Message-Id: <1272928899.24081.62.camel@localhost.localdomain>
tis 2010-04-06 klockan 16:41 +0100 skrev Nathan:

> Is this a change in semantics which the server considers to be
> equivalent (semantically), or should/must the representation equivalence
> be considered from a user-agent / cache perspective.
> 
> Would two representations with different content-encodings and character
> encodings be valid replacements for each other at a cache?

Any change in properties which would make the response incompatible with
different is not semantically equivalent.

It does not matter which axis this property change is on, be it
Content-Encoding, charset, Content-Type, etc.

Only changes which do not change response compatibility can be
considered equivalent, such as changes in compression level, xml
formatting, or even any of the above when done based on other parameters
than the request such as local server configuraton (i.e. default charset
setting when reading files from filesystem).

> "weak validators allow for more efficient caching of equivalent objects"

Yes, as only the first response will get cached and then reused for all
subsequent requests where the serer says the weak etag matches.

Weak etags just need to be unique enough that the server can say "yes,
whatever you have cached earlier with that etag is valid as response for
this other request you sent me now".

Regards
Henrik
Received on Monday, 3 May 2010 23:22:12 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Friday, 27 April 2012 06:51:18 GMT