REPLICATING BACK-END AND FRONT-END APPLICATIONS
Some replication applications will benefit from a back-end/front-end design. In this design, the front end contains the forms, reports, and code; the back end contains the tables and data. The front end uses attached tables to the back-end data. (For more information on front ends and back ends, see Chapter 22, “Handling Multiuser Situations.”)
The front end and back end are two separate replica sets, and each must manage its synchronizations separately.
You can develop a front-end application by using sample back-end data. When the development phase is complete, the application is synchronized with other replicas in the replica set that attach to the “real” back-end data. This development methodology allows third-party developers to distribute applications to numerous customers who never need to share private data.
Note
Linked tables have special attributes in replicated databases. Links to tables created in the Design Master are propagated to all replicas. A linked table is a global property. The actual path to the attached table is also propagated but is treated as a local property. The path created at the Design Master is propagated to replicas as a default. A replica may modify the path to the attached table, even if that replica isn't the Design Master.