I'll have to think about this a bit more, but I can see that the usual TP2 puzzles do not apply in the wave case, even when multiple clients submit changes to the server concurrently as the server ensures that all clients see all changes in a consistent order. The common TP2 puzzles occur when operations are received in different orders at nodes in a peer-to-peer network.
The part I'm not sure about though, is the claim that only a single state space is used by the server and how this relates to when acknowledgements are sent from the server to the client. In the end I'm trying to think about whether the introduction of the server acknowledgement means that latency between clients (ie, how long it takes operations performed on one client to be seen on another) is related to the number of concurrent operations performed by clients. The OT white paper claims that the latency is constant, though it does not specifically talk about concurrent operations by clients when making that claim. Cheers, Dan On Nov 19, 9:17 am, Sam Thorogood <[email protected]> wrote: > Again, not an expert in the academic side of OT: but, the only > simplification we employ (and I suspect this is just a limitation of > our optimistic client) is that each client only has one delta (i.e. > set of operations) 'in-flight' at any one time. However, there's no > limit on the number of submissions from different clients. Each delta > is still applied in some timing-related order: one delta will always > be before or after other deltas, never 'on top'. > > > > On Thu, Nov 19, 2009 at 10:47, Daniel Paull <[email protected]> wrote: > > One thing that is not clear from the Google Wave white paper on OT is > > when acknowledgements are sent from the server to the client. What > > I'd like to determine is if operations are sent from a number of > > clients to the server concurrently, or, does the server make clients > > take turns in sending operations to the server. Though I haven't done > > the math yet, I would expect that if clients send operations > > concurrently to the server and the server maintains only a single > > state space, then the OT transformation functions would need to > > satisfy TP2 (which is avoided in the Jupiter system). If I'm not > > right here, are there any resources around that formally prove > > correctness in the OT approach of Wave that I can read? --~--~---------~--~----~------------~-------~--~----~ You received this message because you are subscribed to the Google Groups "Wave Protocol" group. To post to this group, send email to [email protected] To unsubscribe from this group, send email to [email protected] For more options, visit this group at http://groups.google.com/group/wave-protocol?hl=en -~----------~----~----~----~------~----~------~--~---
