OBJECTS AND EVENTS.

Abstract

This part 3 of a 14 part series on database design explores the Process-Driven Data Design. It gives three reasons to distinguish between objects and events. Events are derived from objects but not vice versa. This means that in constructing a long-range plan - an overall database architecture - for a pool of shared data, we begin with the easily identified object records and then derive events by Bachman manipulation. Second, shared files imply standard data names and, most especially, standard keyfold formats. Both ideas can be understood more easily if we first apply them to involatile reference data, rather than to less widely familiar events. Third, how we physically implement a design into our database managment system depends on each record's volatility in many ways. Multiple access paths, backup/recovery procedures, and migration techniques, all vary with volatility.