Re: Syncing and partial (truncated) results

[Ezra Cooper <ezra@xxxxxxxxxxxx>]

On Oct 23, 2004, at 11:20 AM, Robert Sayre wrote:

> Ezra Cooper wrote:
> > The crux of this is that, since we're working backward in time, no  
> > surprises can come once we begin the sync. Also, since we're using 
> > date  ranges, a series of ranges is guaranteed to pick up everything 
> > in that  range; using ordinal indices, the meaning of a range request 
> > would  change whenever anything changes in the list, leading to 
> > duplicated or  missing items.
>
> OK, I think I understand how this works.
>
> It would be really helpful if you could document an example series of 
> requests and responses. Check out the "Examples" section of 
> PaceServiceError[0] which shows "Interaction of a server with three 
> clients ('Client A','Client B', and 'Client C') over time".

I've written up a (slightly) more precise proposal at 
PaceReturnCollNewestToOldest [1] including a big example. I want to 
particularly note that the representation of the collection is not a 
material part of this pace (we should take that up separately).

A further explanation of my thinking of why this is useful is at 
SyncIsTheParadigm [2].

Ezra

[1] http://intertwingly.net/wiki/pie/PaceReturnCollNewestToOldest
[2] http://intertwingly.net/wiki/pie/SyncIsTheParadigm