Date: Wed, 31 May 2000 07:55:28 -0400 (EDT) Message-Id: <200005311155.HAA25138@tantalum.atria.com> From: "Geoffrey M. Clemm" <geoffrey.clemm@rational.com> To: ietf-dav-versioning@w3.org Subject: Re: Why do we need working resource ids ? From: jamsden@us.ibm.com Here's a couple of compelling use cases: the server generated URLs are not likely to be in DAV namespaces. That is, collection semantics won't work. They're only to access one specific resource. The collection semantics do work with the versioned resource and Target-Selector. A revision id and working resource id don't give you any more collection semantics than a revision URL or working resource URL, so I don't see your point here. The collection semantics apply to the human define versioned resource URL's, and that is true whether or not you use revision URL's or revision id's to identify particular revisions. I am not yet compelled (:-). Second, we need Target-Selector anyway for selecting revisions by label, workspace, activity, perhaps baseline - all of which are logical target selectors that have human meaningful names. I agree that we should have a Target-Selector header that specifies a label. I disagree that it should specify anything else. We have the ability to select revisions by workspace ... just use the URL of that workspace extended by whatever member of the workspace you'd like to refer to (or use a Workspace header and a relative URL). We have agreed (as recently as last week in Seattle :-) that we would not allow an activity or a baseline in a Target-Selector header, but rather require that a client merge that activity or baseline to a workspace and then use that workspace. The reason is that the caching opportunities provided by a workspace are required to allow a server to efficiently perform activity or baseline target selection. We don't want clients to have to use server generated URLs in some instances and versioned resource and Target-Selectors in others in order to use the more meaningful names. Personally, I think Target-Selector will be used 99% of the time with either a label or workspace. The server generated URLs will probably be rarely used. I agree that a Target-Selector should be used to specify a label, and that we have a Workspace header in which you can specify a workspace. The question is whether we need revision-id's and working-resource id's. But your reasoning above, it appears we do not, since 99% of the cases use labels and workspaces, and the (rarely used) server generated URL's can be used for the remaining 1% of the cases. Cheers, Geoff