- From: Mark A. Hale <mark.hale@interwoven.com>
- Date: Tue, 30 Jan 2001 12:16:35 -0800
- To: <ietf-dav-versioning@w3.org>
> - Thanks for catching the 7.5 typo ... will fix. Thanks. > - Concerning the statement about working-resource and > workspace options, I'm happy to delete this sentence, > since I don't see that it improves interoperability, > and could incorrectly lead a reader to believe that > the options are inconsistent. So I'll delete this sentence > unless anyone objects. Thanks. > - As for the complexity of the protocol, if there is > something that you would like to see addressed before > we send the protocol to the IESG for "proposed standard" > status, please identify the specific issue during the > working group last call period (i.e. before 2/1/01). > It is my belief that the complexity issues have been > adequately addressed by clearly identifying the part > of the document that should be read by those only > interested in "core" versioning functionality. Each of > the options consist of functionality that is currently > supported by multiple versioning repositories, and > therefore it significantly contributes to interoperability > to provide an option that standardizes access > to that functionality. I agree with what you've stated here in the segregation of the functionality to aid in understanding and implementation of the current specification. My thoughts about complexity are in two areas. 1) HTTP The marshalling in DeltaV is designed around WebDAV's premise as an HTTP extension. This was a design decision for WebDAV and I am not going to argue on way or the other on this issue. My only comment is that there may be people that want to implement the WebDAV+DeltaV specification that participate in other transport agnostic standards. In that case, DeltaV is a very specific implementation and may appear complex relative to the other agnostic activities. 2) Marshalling IMHO, the marshalling in the current version of the specification is very fine grained. The number of properties is large and the number of client-server transactions to accomplish a specific activity can be potentially as large. My experiences have shown that fine grained specifications can also result in less interoperability. In this case, I speculate that disparate client and server vendors will need to work closely to insure that the implementations are consistent and function as expected on either end. This may result in one-off client-server implementations. Again, this is speculative and does not specifically address items in the proposal. A number of others have also commented on complexity. I am interested in seeing how they describe their thoughts. Overall, the team had worked hard to develop a specification on a very difficult subject and I for one appreciate the efforts that have gone into this. Thank you and have a good day, Mark
Received on Tuesday, 30 January 2001 15:16:39 UTC