- From: <Tim_Ellison@uk.ibm.com>
- Date: Tue, 6 Mar 2001 16:24:56 +0000
- To: ietf-dav-versioning@w3.org
Steve, To be precise, an activity selects a set of resources, it does not track the methods applied to the resources. That is to say, while it may be a convenient short-hand to add a DAV:activity to VERSION-CONTROL it is not capturing the fact that 'a versionable resource was version controlled', but rather 'this version is associated with this activity'. In fact, my answer would have been that you should PROPPATCH the DAV:activity-set property of the root version created by the VERSION-CONTROL -- but I do not object to the shorthand extension to the VERSION-CONTROL method. Using activities to track changes to a resource requires that you reconstruct the changes that took place based on the set of resources captured by the activity. Regards, Tim Ellison Java Technology Centre, MP146 IBM UK Laboratory, Hursley Park, Winchester, UK. SO21 2JN tel: +44 (0)1962 819872 internal: 249872 MOBx: 270452 |--------+------------------------> | | "Steve K | | | Speicher" | | | <sspeiche@us.i| | | bm.com> | | | | | | 2001-03-06 | | | 03:28 PM | | | Please respond| | | to "Steve K | | | Speicher" | | | | |--------+------------------------> >--------------------------------------------------------------------| | | | To: ietf-dav-versioning@w3.org | | cc: | | Subject: Re: How to supply an activity for | | VERSION-CONTROL? | | | | | >--------------------------------------------------------------------| Jim, It would seem that if someone would want to track all the changes for a given activity, that the initial version would be part of those tracked changes. I would think it would be a problem for someone who would want to see what the "reason" was to bring the activity into existence. The figure in section 11 shows the resource "foo.html" with version "V1" associated with activity "Act-1". How is it possible to have this situation if VERSION-CONTROL doesn't accept the DAV:activity property as described by Geoff? It seems that in order to maintain the integrity of controlled resources, a server may want to restrict a given namespace to require an activity on all updates. Why shouldn't the protocol support this? I realize that this could render many non-activity aware clients useless. Thanks, Steve Speicher sspeiche@us.ibm.com
Received on Tuesday, 6 March 2001 11:25:55 UTC