- From: Clemm, Geoff <gclemm@rational.com>
- Date: Thu, 1 Nov 2001 14:42:50 -0500
- To: ietf-dav-versioning@w3.org
From: Peter Raymond [mailto:Peter.Raymond@merant.com] "12.3.1 DAV:baseline-collection (protected) This property contains a server-defined URL for a collection, where each member of this collection MUST either be a version-controlled resource with the same DAV:checked-in version and relative name as a version-controlled member of the baseline-controlled collection at the time the baseline was created, or be a collection needed to provide the relative name for a version-controlled resource." This seems to be saying that live properties like DAV:displayname will not be captured by the baseline (only DAV:checked-in will be captured). The semantics of DAV:checked-in require that the dead properties on the VCR are identical to the dead properties on the DAV:checked-in version, but you are correct that this requirement does not apply to live properties. But the protocol does not prevent a server from capturing live properties, it just does not require it to do so. In the case of D:displayname, I expect that most servers will chose to capture that as part of the version state, but that is up to the server. In my opinion a baseline should record the exact state of the configuration (including properties of it's members). A VCR must have the same content and dead properties as it's checked-in version, but it's live properties may differ and would not seem to be captured in the baseline. All the reasons for why we didn't require live properties to be captured by versions (I'm not sure if that issue made it from the mailing list into the FAQ yet) apply to why would wouldn't require them to be captured by baselines. (examples include the DAV:version-controlled-configuration property and the DAV:lockdiscovery property). Is my understanding of this correct? Yes. Do other members of the group agree that it makes sense to capture other live properties of the VCR that is under baseline-control? I agree that a server should be allowed to capture other live properties, but I do not agree that we could/should require that all live properties be captured by a version/baseline. Cheers, Geoff
Received on Thursday, 1 November 2001 14:43:29 UTC