Re: Why MD5 Headers are Imperative

Eric J. Bowman schrieb:
> Here's the server-side resource, no "representation" to it, this is the whole 
> enchilada:
>
> http://canuck.bisonsystems.net/dev/atom/entry-server-2.xml
>
> Here's the client-side, writable representation retrieved by an APP client:
>
> http://canuck.bisonsystems.net/dev/atom/entry-server-2.xml
>
> Here's the read-only representation with eTag=atom:id+atom:updated for the 
> majority of use cases, which are read-only and do not even see the <p/> tag:
>
> http://canuck.bisonsystems.net/dev/atom/entry-server-3.xml
>
> In my implementation, editors can move that <p/> tag around all they want 
> without altering the published read-only representation, provided an Atom-
> friendly publishing protocol...

To me this means that you've implemented a publishing system with 
staging, and that there are two different resources. Give them different 
URIs, and the caching problems should go away...

> ...

Best regards, Julian