Re: MKRESOURCE results
Geoffrey M. Clemm (geoffrey.clemm@rational.com)
Mon, 20 Dec 1999 17:01:23 -0500
Date: Mon, 20 Dec 1999 17:01:23 -0500
Message-Id: <9912202201.AA07785@tantalum>
From: "Geoffrey M. Clemm" <geoffrey.clemm@rational.com>
To: ietf-dav-versioning@w3.org
In-Reply-To: <1999Dec20.133835.1250.1423429@otismtp.ott.oti.com>
Subject: Re: MKRESOURCE results
From: Tim_Ellison@oti.com (Tim Ellison OTT)
I'd like to open the debate (if anyone is interested :-) into the effect of
MKRESOURCE for a versionable resource.
Presently, if a versionable resource is created in a versioned collection,
the created resource is necessarily versioned. Presumably if the same type
is created outside a versioned collection it is not versioned.
Yes.
When the new resource is versioned what is the state of the resource? Is
there one revision and no working resources (i.e., requiring MKRESOURCE,
CHECKOUT, PUT, CHECKIN) or is there no revision and one working resource.
I prefer a third alternative: one revision (with default values for all
properties and an empty body) and one working resource (checked out from
that revision. It is the working resource that the PUT or MKRESOURCE is
applied to.
At the moment the protocol allows setting properties via MKRESOURCE so that
sufficies for some cases where the resource can just be created in it's
final form; however, configurations are an example where subsequent,
immediate updates will be most likely.
Yes, I definitely think the result of the PUT or MKRESOURCE should be
a working resource, not a revision.
Would having a working resource with no revisions be a bad idea?
It require clients to special case such working resources (i.e. handle
the fact that there is no associated predecessor), and would open up the
question of what is left after you "uncheckout" that working resource.
Overall, I think it is cleaner/simpler to just have an empty initial
revision, and a checked out working resource.