Re: Revision names

Chris Kaler (ckaler@Exchange.Microsoft.com)
Tue, 12 Oct 1999 15:02:43 -0700 

Message-ID: <FD7A762E588AD211A7BC00805FFEA54B041DD98A@HYDRANT>
From: "Chris Kaler (Exchange)" <ckaler@Exchange.Microsoft.com>
To: "'Tim_Ellison@oti.com'" <Tim_Ellison@oti.com>, ietf-dav-versioning@w3.org
Date: Tue, 12 Oct 1999 15:02:43 -0700
Subject: RE: Revision names

  <chris/>
  2) Not being URIs, there are conflicts with cross server

  <tim/>
  I don't understand this.  Using "revisionid:42" makes it no more or less 
  distributable than "42" alone.

<chris2/> URIs are supposed to be globally unique...

  <chris/>
  3) You limit the stores ability to create its own URLs

  <tim/> servers would have to produce conforming URIs

<chris2/> Which means that they can use their own format...

  <chris/>
  I don't understand your point about REPORT.  You can report
  multiple revisions of the same resource regardless of how
  you represent the URL.

  <tim/>
  My point was that unless you had a URI to a specific revision, you could
not 
  issue a REPORT request that selected two revisions of the same resource 
  since there is currently only a single Target-Selector.

<chris2/> I disagree.  
            <conflictitem>
               <href>...</href>
               <ref>...</rev>
            </conflictitem>
          However, I do want to have revision-specific URLs, I just want
          the server to be able to give them ANY value

  <tim/>
  I agree that the client would not be able to parse the URL in general.
  How should clients get the revision-specific URL?
<chris2/> There is a property -- well there was a property -- it seems
          to be gone now???  There was a revisionurl property on revisions.
          Geoff -- when/why did we remove this?