Re: Branching Scenarios

Geoffrey M. Clemm (gclemm@tantalum.atria.com)
Sun, 21 Feb 1999 14:05:04 -0500


Date: Sun, 21 Feb 1999 14:05:04 -0500
Message-Id: <9902211905.AA03572@tantalum>
From: "Geoffrey M. Clemm" <gclemm@tantalum.atria.com>
To: ietf-dav-versioning@w3.org
In-Reply-To: <9901199194.AA919487591@SMTPGwy.MERANT.com>
Subject: Re: Branching Scenarios

   Author:  jamsden@us.ibm.com at SMTPPOST

   The more I think about it, the more I believe the protocol should not be 
   involved in specifying policies for
   specific repository implementations or legacy data integration. This seems 
   like something that would be impossible to describe, and impossible to even 
   determine what arbitrary repository managers might want to do. I would 
   suggest this is a server implementation issue, not a protocol issue.

I take an intermediate position on this issue.

I do believe that handling the needs of branch-based systems are
essential, and that it is important to verify this by mapping these
needs into a canonical set of scenarios and properties for use by
branch-based clients.

But I believe it is important that these properties and scenarios
should *not* be in the base protocol, but rather should be in a
separate document (such as is being done for the proposed structured
document extended properties).  As Jim indicates, there are a huge number
of special policies and implementations found in each of the various
CM implementations, and the base protocol would be seriously harmed if
we weighed it down with all the detailed scenarios and properties
that are relevant to only a certain kind of server implementation.

Cheers,
Geoff