I would like some feedback on using user defined types to store
business logic in Oracle. For instance all business logic for say a
User entity would be placed in a User Defined Type. Modification of
such data would be accomplished by instantiating the "User object" and
using getter and setter methods, simillar to other OO languages. A save
method would be used for the Inserts, Updates, Deletes. This would
allow a check rules method where business logic could be checked before
updates for instance, regex on email addresses, phone numbers, checking
priviledges of the user. If another application were to need to access
and manipulate the user object in a different way, a derived object
could be created overriding rule checking of its super object. I am
familliar with the caveats of OO pl/sql, like true encapsulation, and
limits on datatypes, however the lure of inheriting objects and
changing methods to suit the needs is very appealing. I have worked
with applications where business logic resides in packages, these
packages over time seem to get very bloated with conditional logic, if
this then that but only when so on... Objects would allow changes using
the open closed principal and therfore inheritance would be used for
new functionality and not break existing code, as changing long
procedures often does. Our organization is commited to Oracle so
database independance will never be an issue, a third tier app server
would just get in the way, our environment will always be homogenous.
Anybody tried this, any reasons this is a bad idea, all feedback
appreciated.
--Chris
Received on Thu Feb 09 2006 - 09:23:52 CST
