- From: Geoffrey M. Clemm <geoffrey.clemm@rational.com>
- Date: Fri, 9 Feb 2001 21:22:44 -0500 (EST)
- To: ietf-dav-versioning@w3.org
From: "Boris Bokowski/OTT/OTI" <Boris_Bokowski@oti.com> section 1.2: Word problem: error - reference source not found Will fix (so Word guru's: any way to ask Word whether there are any bad references?). section 2.1, third paragraph: mention that the new version resource is a root version resource. Some versioning systems allocate a dummy (null) root version, in which case the new version would be a descendant of that null root version. section 2.2.3, 2.2.4: DAV:supported-method-set and DAV:supported-live-property-set should be protected properties. Yup! Will fix. (Will also fix DAV:supported-property-set). section 2.3.2: add: "This property is removed when the resource is checked in, and then added back (potentially identifying a different version) when the resource is checked out again." Good suggestion. Will do. section 2.5, postconditions: change "a new version history is created and a new version resource is created" to "a new version history is created and a new root version resource is created" (add "root") See comment on 2.1. It might not be the root version. section 7.2.1: "If the resource is associated with a workspace": when is a resource associated with a workspace? If one of its parents is a workspace? When the parent has a DAV:workspace property, then the resource must have one as well, but a server may decide to associate other resources with the workspace (this was added for Mark, who wanted to be able to have his server associate working resources with a workspace). section 7.7: missing precondition: parent of null resource MUST exist (or is this covered elsewhere?) Yes, that is required by RFC 2518. postcondition: add "and whose workspace property identifies the containing workspace" We could add this postcondition to every method that creates a resource, but I'd prefer to just state it as an axiom in the DAV:workspace property definition. This keeps the requirement localized to the workspace option. section 8.3.1, last sentence: change "was" to "were to be" Will do. section 9.2, Preconditions: DAV:must-select-version-in-same-history is misleading; you cannot select a version from another history, but you could select no version if the label was not found. The precondition should be DAV:must-select-version-in-history. (This applies to other sections listing the same precondition name, e.g. 9.6) Good point. Will do. section 9.8: This section only applies when the server supports the WORKING-RESOURCE option Well, I'd say that the section applies always, but the postcondition cannot be satisfied unless the server supports the working-resource option. section 9.9,Preconditions: DAV:apply-to-version: where is this defined? Yes, Greg noticed this too ... this precondition has been fixed and moved to CHECKOUT. DAV:label: should this be DAV:label-name ? Good catch! section 10.2.2 URLs That's always a hard call. I have consistently used "URL's" instead of "URLs" in the document, since URLs can sometimes be confused as a four letter acronym. I could change this if anyone cares enough to support this change request (otherwise it's just 1-1, and the author wins all tied trivial formatting issues :-). section 10.4.1 should read: "This property identifies the members of the workspace collection that are collections under baseline control." also: Can't the workspace itself show up in this list too? Yup. Will fix. section 10.5, third paragraph: change "members" to "member" I'll rewrite this paragraph ... the postconditions say that the new baseline is empty, and that's what this paragraph should say as well. section 10.5, Postcondition DAV:select-existing-baseline: How does this deal with pre-existing non version-controlled resources in the collection with the same relative name? The preconditions only disallow pre-existing version-controlled resources. Good point. I'll change that to "must have no members". section 10.10, Postconditions: baseline-controlled-collection -> DAV:baseline-controlled-collection-set Actually, baseline-controlled-collection is right. A version-controlled configuration has a DAV:baseline-controlled-collection, not a baseline-controlled-collection-set (only workspaces have the latter). Boris: Please thank your colleague for this excellent review! Cheers, Geoff
Received on Friday, 9 February 2001 21:24:17 UTC