Re: i22: ETags on PUT responses

On mån, 2008-01-07 at 14:24 +0100, Julian Reschke wrote:

> > It's also not clear how to handle ETag then a client makes a transition
> >>from Request-URI to Content-Location URI for the purpose of authoring a
> > negotiated resource or similar.
> 
> Good point. That would require a new HTTP feature, right?

Not sure. I don't think so. Just clarification I guess.

The natural thing for most implementers I guess is to assume equalivence
between the two locations (Request-URI and Content-Location) when moving
from Request-URI to a direct Content-Location reference, but the correct
thing here is to discard any response received from the negotiated
Request-URI when moving to the direct Content-Location URI, and only at
most use ETag as a "weak" comparision locally in the client to verify
that the new response makes sense without using it in any HTTP
conditions.


But yes, there may be a need for a HTTP extension defining reliable URI
namespaces, but that's probably outside our charter and more a topic for
possibly the next generation WebDAV trying to do things more inline with
HTTP.. 

  - How to express negotiated resources better
  - How to express when a resource exists at multiple URIs
  - Derived/related resources such as properties of the resource
  - And to fit all this into both addressability and cache
invalidations..

Regards
Henrik