posted 21 years ago
Yup, for CMP the correspondence is going to be close. Ignoring other ways of getting data (views, stored procedures), what you'll have is:
T = number of tables we want to provide data to/from the application
R = number of relationship tables
O = number of other tables maintained outside the application
(e.g. via triggers)
T + R + O = the total number of tables
So "T" would be the number of CMP-based entity beans you'll have. The "R" tables are used for maintaining the constraints, and they'll be mentioned in whatever container-specific deployment descriptor or tool you use for specifying your CMR implementation. The "O" tables are just stuff you are required to have but the application may not care about, like automatically maintained audit trails.
Reid - SCJP2 (April 2002)