On Thu, May 27, 2010 at 10:09 AM, Tad Glines <[email protected]> wrote:
> This is incorrect. Remember that a delta may contain operations on more than
> document within a wavelet. In the case of the FedOne server, all operations
> contained in a delta are applied in order without first trying to compose
> all operations destined for a particular document within the wavelet.
>

I had forgotten about multiple documents. There's no reason that a
server can't compose the operations before applying them, however, so
that should stand as a caveat.

> Also, remember that all deltas MUST pass through the federation host (for
> transformation) before being send to clients (as an update). Because the
> "server" serializes (processes them one at a time) all deltas for a wavelet,
> the wavelet history at the server and at the client (at quiescence) will be
> identical.
>

No, the wavelet history will not be identical, the final state will be
identical. Unless you're using an extremely primitive client like the
FedOne example which does not maintain its own state internally. Even
in that case they'll be seeing different histories if they're talking
to different servers. (The whole point of OT is that different
sequences of deltas converge on the same final state).

-- 
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.

Reply via email to