Next message: Geoffrey M. Clemm: "Re: "stable" href's"
Date: Thu, 27 Jan 2000 00:19:07 -0500
Message-Id: <10001270519.AA29142@tantalum>
From: "Geoffrey M. Clemm" <geoffrey.clemm@rational.com>
To: ietf-dav-versioning@w3.org
Subject: Re: DAV:revision-resourcetype
From: jamsden@us.ibm.com
I never thought having MKRESOURCE create a versioned resource was
necessary. The scenarios in the goals document talk about creating a
resource and getting it in some reasonable initial state before committing
it to versioning. The base case is create an initial resource with
MKRESOURCE and then version it with VERSION.
The main use of DAV:revision-resourcetype is to tell the server how to
store the revisions of that versioned resource. For a simple
resourcetype (like text/xml), there might be several "storage types"
(e.g. compressed, text-delta, binary-delta).
So trying to infer the DAV:revision-resourcetype from the
DAV:resourcetype of the first revision will not always be sufficient.
Having MKRESOURCE create
versioned resources seems like unnecessary method overload. The fact that
we are having trouble determining the resource type is an indication of the
consequences of such an overload.
I'll take this opportunity to modify a statement I made in an earlier
message. Instead of: "the DAV:resourcetype of each revision of a
versioned resource must be the same as the DAV:revision-resourcetype
of the versioned resource", I should have said: "the DAV:resourcetype
of each revision of a versioned resource must be compatible with the
DAV:revision-resourcetype of the versioned resource".
Cheers,
Geoff