Re: Revision identifier and revisions la
jamsden@us.ibm.com
Tue, 5 Oct 1999 13:44:23 -0400
From: jamsden@us.ibm.com
To: ietf-dav-versioning@w3.org
Message-ID: <85256801.00619618.00@d54mta03.raleigh.ibm.com>
Date: Tue, 5 Oct 1999 13:44:23 -0400
Subject: RE: Revision identifier and revisions la
<jm>
confused me with the () part. "Revision names" are not defined to have
any particular semantics except for being revision-ids or labels (at
least I could not find it in the spec). Labels and revision-ids are
independently defined to be unique for a given resource. If "revision
name" did have the semantics you suggest, then the id and label
namespaces would coincide and we would have a problem.
I propose that we systematically erradicate "revision name" from the
spec. It adds no semantics and is confusing.
</jm>
<jra>
The revision names for a revision are unique. There are two kinds of revision
names, server supplied (revision id) and user supplied (revision label). The
implicatation is that they share the same namespace. The reason for having the
revision id is that a revision must have at least one revision name in order to
distinguish it from any other revision. Since this is required, the server
should supply it. Users can then add others as needed.
</jra>