Re: Web and Mail integration: a few key connections.

• Mail folder: WWW Talk Oct 93-present
• Next message: ts: "Oracle and <FORM>"
• Previous message: Dave_Raggett: "Re: Forms submission"
• In-reply-to: Keith Moore: "Re: Web and Mail integration: a few key connections. "

Date: Mon, 8 Nov 93 12:03:19 +0100
From: Tim Berners-Lee <timbl@www3.cern.ch>
Message-id: <9311081103.AA07661@www3.cern.ch>
To: Keith Moore <moore@cs.utk.edu>
Subject: Re: Web and Mail integration: a few key connections. 
Cc: iiir@merit.edu, uri@bunyip.com, www-talk@nxoc01.cern.ch,
        ietf-822@dimacs.rutgers.edu, moore@cs.utk.edu
Reply-To: timbl@nxoc01.cern.ch

>There is a need for a URN-like thing that can be absoultely tied to  
a
>particular format of an object.  Without such a URN there is no way  
to
>cache/replicate objects correctly. 


(URN-like thing =~ "URI", the generic identifier)

The HTTP spec has, in the object metainfo, a URI: header.
It specifies a URI for the object.
The "vary=" parameter of this header specifies whether the
URI refers to the exact binary pattern, or whether it can
refer to "variants" of language, content-type or version.
(These are the only variants curerntly explicit, one can
imagine others)
Caching systems obviously use this information intelligently
to know whether they have a copy verbatim of an object or a copy of
one instance of a generic object.

Tim