I think ProofOfLife has pretty much covered the issue of triggers, and I agree with what he has said. Triggers are difficult to test and debug so I generally avoid them unless I am trying to enforce a complex business rule at the database engine level. I also seldom use the ADP approach as Microsoft no longer is enhancing the functionallity - but they apparently will still support it through at least the next version. I agree with the issue of storing a value that can be calculated - I usually avoid that approach and calculate the value when I need it.

On the issue of replication and triggers, I can shed some light. Triggers will fire anytime the data is updated - presuming you use an UPDATE trigger. However replication does add a layer of complexity, and depending on the number of replicas, a fair bit of administrative overhead. Are the remote sites able to use a web app to do updates? Both require Internet connectivity, although the replicas do allow people to enter data when the link is down, at least for periods of up to 48 hours.

Wendell