- From: Greg Stein <gstein@lyra.org>
- Date: Tue, 16 Jan 2001 11:05:06 -0800
- To: ietf-dav-versioning@w3.org
On Tue, Jan 16, 2001 at 01:33:47PM -0500, Geoffrey M. Clemm wrote: >... > In subversion, the BASELINE-CONTROL is a trivial operation, because > your baselines are already in traversable form. So the cost to you is > that protocol asks you to "publish" important baselines in the form of > baseline-controlled collections, i.e. give them human meaningful > URL's. > > I agree that the intermediate BASELINE-CONTROL step is not optimized > towards your implementation (which does not need it), but it is > needed for interoperability with most other baseline implementations. That is exactly the issue that I'm pondering on :-) Okay... thanks for the feedback. I'll figure out a way to deal with this, then. I don't think that we can be truly interoperable because of the question "where can I create that (temporary) baseline-controlled collection (BCC)?" I can detect an SVN server and know where an automatically constructed BCC exists, but if the server is *not* an SVN server, then I will need to know a location to create that BCC. [ effectively, I will automatically create a BCC each time a baseline is created (via a check-in/merge). a property on the public VCR will tell me the collection containing the BCCs. ] [ attaching labels to the baselines is related to this issue ] Cheers, -g -- Greg Stein, http://www.lyra.org/
Received on Tuesday, 16 January 2001 14:05:21 UTC