Designing Client/Server Applications › Types of Logic › Data Manipulation Logic

Data Manipulation Logic

A key design consideration concerning data is how important is access to current data. Depending on the application and business needs, data that is updated periodically, instead of continually, may be adequate. Data that is replicated on the client platform is often used only for read access (data look up). If other actions are required, such as create, update, and delete, the database management system (DBMS) must have the capability to handle the data integrity, node directories, and two-phase commits. Typically, the DBMS provides these capabilities and not the generated application. You can, however, build with CA Gen the time stamping and locking logic in your application, but you must also design for various failure conditions.

With read-only replicated data, you have several design decisions:

• How to distribute the data
• How to update the replicated data
• How often to update the replicated data and ensure that each client machine has the same data
• How to ensure consistency and synchronization of data after a component fails

You provide these capabilities through the processing logic in your procedure action diagrams (PrADs), the procedure steps on the Dialog Flow Diagram (DLG), batch transfer of data, and office procedures.