Serge wrote:

>
> Dalton,
>
> Is there any way to take a look at source code, docs (if any)? Probably we can try to resurrect Chrisitian's idea and create open source project?
>
> Thanks,
> Serge
>

Hi Serge,

Most of the code is very specific to use (and also belongs to
distributel...).
The concepts of how to log, what to log, how to reconsile the logs and
what constitutes replication (concurrancy, load balancing, client side
transaction caching, fail over concurrancy) all affects the project and
how the code is implemented.
We have attacked each area, and they are not simple subjects to explore.
With Jasons response that he is already producing a commercial version
of the 'IBOTS' that I use so frequently, I do not want to discuss the
issue unless he says that he wants it discussed.
The one thing I do realize is, that alot of users do not understand what
the term 'replication' means. It could be as simple as a live database
being updated from a series of transaction logs (ala oracle style
replication/backups) to having the ability of pulling the plug on a
working server, and the client reconnects to another server, checks to
see how much of it's work from this session made it into the backup
server, and applying the differences. (with the associated
reconsiliation so that the data is not inserted twice due to the IBOTS
also trying to move the data).
The concept of having databases go off-network (like in a laptop)
operate for a while and then merge up with the main database at a latter
time, also constitute a form of replication.

Just a few random thoughts

best regards

Dalton