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

Jamie Lokier wrote:
 Julian Reschke wrote:
>> On 06.04.2010 16:49, Jamie Lokier wrote:
>>> ...
>>> So back to an earlier query: When *would* you use weak etag
>>> equivalence for different representations?  If never, do weak etags
>>> have any purpose at all?  What would you use weak etags for?  Because
>>> if you only use the same weak etag when representations are identical,
>>> you should be using strong etags instead for that.
>>> ...
>> For instance, the way you construct your representation may not 
>> guarantee binary identity, although the underlying resource is the same 
>> (think an XML database with a generic XML serializer, affecting, for 
>> instance, attribute order).
> 
> Well, Yves Lafon said: 
>>> No reason to forbid it, but if you rely on weak etag equivalence for
>>> different representations, expect it will break.
> 
> Changing attribute order *is* making a different representation.
> 
> Thus Julian's idea of when to use the same weak etags is recommended
> against by Yves.
> 
> I happen to agree with Julian(*), but I'm really interested in when
> Yves thinks it is appropriate to use same weak etags, given the
> statement that it's unsafe to use them for different representations.
> 
> -- Jamie
> 
> ps. (*) Why my pov matches Julian's:
> 
> Instead of changing attribute order, what about different output
> encodings at serialization time?  Note that the *input* XML database
> might not even *have* a character encoding.  I think that is in the
> same category as attribute order, because the underlying XML resource,
> right down to the individual characters, is unchanged, although it's
> clearly not binary identical.
> 
> If the underlying resource is the same for different character
> encodings (Content-Type charset=), in what way is compression
> (Content-Encoding) semantically different?  I think it isn't.

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?

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

regards

Received on Tuesday, 6 April 2010 15:42:10 UTC