How to Handle Persistent Objects?

There are a variety of ways to handle persistent object storage:

• Data Base Management Systems
  DBMS are an excellent solution if your data model is complex; many designers use a DBMS unnecessarily and cause their users stress. For simple data models, consider a Flat File system which puts all the records in a single table.
  

• Text/Binary Files
  Files don't have a record structure; the structure is implied in the code. Files are popular because the software is inexpensive and the performance is good for some types of applications.

• Access Methods
  This is a fancy word for file systems that support a record structure. Wintel and Unix machines don't have access methods; IBM machines do.

• Streamable objects saved externally
  The program explicitly manages object storage and retrieval, usually to a flat file system.

• Persistent classes
  The class manages its own file structure allowing the program to save and restore objects from disk. This is the proscribed method in CSC 205; design your software so all persistent data is handled by the appropriate classes.

• Distributed Objects
  Many software systems are being written on a distributed, heterogeneous platform. Such systems need a standard way to store and retrieve objects. The Object Management Group (OMG) has specificed an architecture (CORBA) for managing and storing objects in a distributed environment. Microsoft has created their own standard for distributed objects (COM).