W3C home > Mailing lists > Public > w3c-dist-auth@w3.org > January to March 1999

Re: [long] Re: I-D ACTION:draft-ietf-webdav-versioning-01.txt

From: Geoffrey M. Clemm <gclemm@tantalum.atria.com>
Date: Thu, 11 Mar 1999 18:26:28 -0500
Message-Id: <9903112326.AA16110@tantalum>
To: masinter@parc.xerox.com
Cc: francis@ecal.com, w3c-dist-auth@w3.org
   From: "Larry Masinter" <masinter@parc.xerox.com>

   > > > If we defined a collection that contains all revisions of a versioned
   > > > resource, where a member of that collection is named by its
   > > > revision-id, then it would be within these guidelines (i.e. it would
   > > > not be URL-munging) for a client to take the collection URL, extend it
   > > > with a revision-id, and use the resulting URL to locate the desired
   > > > revision.
   > >
   > > No, I think that doing so would be an extension of the currently
   > > known required URL-munging methods.
   > I'm not sure I understand why this would be so.  It seems to me that all
   > you've got here is a collection that is defined to contain references to
   > revisions; the munging going on is the same munging that happens when you
   > ask for a member of a collection.  No?

   In DAV, the members of a collection are designated with complete
   URLs. If the "revision-ID" is used to construct a relative URL which
   is combined with the "version collection" to create the URL of a particular
   version, then you wouldn't be doing any new URL munging, since you'd
   be doing relative URL calculation.

Good (so at least I don't have to redesign that part of the proposed
protocol :-).

   However, you'd want to make sure
   that the relative URLs constructed didn't contain "/" or "." or reserved

I assume you could achieve this by requiring that a client escape
reserved characters in revision-ID's when using them to construct the URL?

Received on Thursday, 11 March 1999 18:26:33 UTC

This archive was generated by hypermail 2.4.0 : Friday, 17 January 2020 20:01:16 UTC