So I suppose you would need an application with relatively stable
btrees for this approach to be effective.

Have you run across a good solution/approach to finding the Nth key
(where N is potentially quite large) in a btree? Or is Jesus' idea of
a second btree of the key positions the best way to go? O(N) is not
bad for updates but large values of N would still seem to be painful
for users in the real world.

Dave

On Nov 19, 2007, at 2:51 PM, David Binger wrote:

> I think the counted btree has a problem in this context because
> every insertion or deletion causes changes in bnodes up to the
> root of the btree.  Every change will be bigger because of this,
> and the probability of a conflict will be much higher because
> the root node is changing every time the content of the tree
> changes.
>
>
>
> _______________________________________________
> Durus-users mailing list
> Durus-users@mems-exchange.org
> http://mail.mems-exchange.org/mailman/listinfo/durus-users
>