Persistence of Data
Self Persisting Classes, page 13
You've seen how OLE Structured Storage allows each application to be responsible for its own objects within a single document. You can use the same technique within your own application by creating self persisting classes.
Say you create an application that contains a number of different classes and you wish to persist collections of objects from each class. One technique involves creating high level routines for saving and loading all of the objects. The disadvantage of this approach is that any changes to a class require that you modify the high level routines. This includes adding new classes, or changing the formats that a class uses to save its data. This kind of centralized approach is necessary in most cases where you are creating a private file format that you are managing yourself.
OLE structured storage makes possible a better approach in which each class knows how to transfer its data to and from a stream. In this case, all the high level routine has to do is call a "Save" or "Load" method for each instance of each class. If a class needs to change its data format, it can do so at will (assuming it maintains the ability to load the older format). The high level persistence routines do not need to be modified in any way.
Use of self-persisting classes can improve the long term reliability of an application by making it less likely that the persistence operation of one class might interfere with those of others used by the application - even when the code or data formats for individual classes are modified.
