- From: Clemm, Geoff <gclemm@rational.com>
- Date: Thu, 21 Jun 2001 00:03:56 -0400
- To: ietf-dav-versioning@w3.org
From: Rick Rupp [mailto:rick.rupp@merant.com] The idea proposed by Tim concerning a new CHECKIN option to replace what is trying to be done with DAV:checked-out-vcr seems good. This allows a client to request the server to do the checkin/update operations and not require it to be done. I'm not sure what you mean by "not require it to be done". I am not convinced that servers supporting other core versioning features should be required in all cases to update the VCR on check in. A server that supports the core-versioning package is not required to support the working-resource feature, and therefore is not affected by this update-the-VCR-on-checkin issue. If the DAV:checked-out-vcr property is made part of the specification then I would prefer it to be OPTIONAL and not REQUIRED. Optional features are close to useless for an interoperable client writer. It's hard enough to write a client that works against the 5 different packages that have been defined, so the only clients likely to take advantage of "optional" features will be those written against a specific server, and those clients don't need the feature defined in the protocol. I'm also wondering if a new core versioning package shouldn't be described similar to the basic-client-workspace package without the update and label features. I'd be strongly against introducing a new package. We already have 5 packages defined, and that is 4 too many. A poor client writer is already faced with too many server variations (and that's just in the versioning axis!). Cheers, Geoff
Received on Wednesday, 20 June 2001 23:58:19 UTC