Data, Processes, and Events: Keys to Comprehensive Analysis - High-Performance Web Databases: Design, Development, and Deployment - page 53

Exhibit 4-7. Entity-life history diagram of the BAD CUSTOMER entity.

Time is another common trigger of state changes. Often, there are
daily, weekly, monthly, and end-of-year processing triggers. Such temporal
triggers are obvious. Some time-based events are less obvious. For exam-
ple, in Exhibit 5 there is a relationship between CUSTOMER and BAD CUS-
TOMER. This relationship is also illustrated in the ELH diagram for BAD
CUSTOMER (see Exhibit 7). The creation of a new BAD CUSTOMER instance
is triggered by time. A customer who does not pay a bill in a certain amount
of time is classified a bad customer — a time-triggered event.

2. Time.

Often, a stage change in one entity is a
direct result of a stage change in another entity. The nine examples of tight
coupling illustrate the ways in which this can happen.

3. Stage Change in Another Entity.

4. Stage Change in the Same Entity.

In many cases, one change of stage in an
entity can automatically trigger another in the same entity. In Exhibit 6, the
creation of a new instance of ORDER can trigger the stage Fulfill Order. The
fulfilling of the order can trigger the picking of the inventory for the order.
The picking stage cannot be triggered by only fulfilling an order but also by
time or an outside influence (i.e., someone writing the picking tickets).
Often, an event occurs because of a combination of triggers.

TOWARD A MORE POWERFUL TOOL SET

The nine types of tight coupling and four types of events are used to for-
mulate a new approach to requirements gathering and analysis. Analysis
can be divided into two major steps: requirements gathering and require-
ments validation. There has been much progress in the past few years in
the validation of requirements. Requirements validation includes such
steps as refinement and normalization of the data model, horizontal and
vertical balancing of the process model, and rigorous synchronization of
both models.

The requirements gathering phase, however, has experienced less
progress. Certainly, Joint Application Design and Joint Requirements Defi-
nition techniques have improved efficiency. These behavioral approaches,