Re: Definition of the word "business" terms?

By supporting views and different levels of presentation, there are ER style modeling tools that can be used to both reverse engineer existing applications and forward engineer them. However, without the capabilities of a robust repository, there are business rules that don't make sense to be part of a data model.

I've actually seen two definitions of ODBMS: a tool used to store objects persitently (for example, an OOPL like Ada with extensions to handle persistence and concurrency control) and a tool used to store object data (for example, an RDBMS with extensions to handle complex and user defined data types with encapsulated function to act upon those data types). I suppose the latter definition would be what you would call an ORDBMS and probably is what you would be interested in most if you were transitioning from relational data to an object environment.

The bottom line is that the RDBMS and ODBMS paradigms are very different. Because the focus is on the data management facet of application development rather than the complete system, the ERD/RDBMS method has inherent limitations in the types of business rules that can be modeled and implemented effectively (by the DBMS).

I have a draft copy of the IEEE IDEF1X-97 standard which contains an extended version of the old purely ER style IDEF1X data modeling language. The document also clarifies the similarities and differences between the two languages and thus could serve as a guide for transitioning to OO. "Identity-style IDEF1X", the OO version, is also designed to be automatable (though of course no tools support it yet).

Eric Bohn Received on Fri Jan 02 1998 - 00:00:00 CST