Data, Processes, and Events: Keys to Comprehensive Analysis - High-Performance Web Databases: Design, Development, and Deployment - page 61

that event. The data stores that represent those entities should also
be attached to the process.
• Each ELH stage triggered by input should have an input-user view as-
sociated with the corresponding processes. For each stage triggered
by input, an analyst should review that stage's corresponding process-
es in the DFD. One or more of the processes should be associated with
an input-user view, which is a data flow going from a process to an ex-
ternal entity. If the stage is triggered by input, its process should indi-
cate that it is receiving input from the appropriate external entity.
• Each stage change triggered by a stage change in another entity
should be represented in the ERD as a relationship between the two
entities. A stage change caused by another stage change is a type of
tight coupling, which is a unique kind of relationship. Thus, one entity
cannot affect another entity unless the two are related.
• In the ELH diagram, each entity must have at least one create stage
and at least one delete stage. An occurrence of any entity must first be
created if it is to be used in a meaningful way. For most practical busi-
ness purposes, information is not kept indefinitely. Even information
that is archived is usually deleted at one point. Thus, all ELH dia-
grams, except those for historical or archival data entities, should
end in a delete stage.

EXPANDING THE COVERAGE OF THE DIAGRAMS

The ELH diagrams — the one in Exhibit 12 for example —
are effective at showing the events affecting a single entity. However, it does
not show graphically the triggers for a given event. It can be enhanced (see
Exhibit 14) by using the trigger symbols used in event network diagrams.

The ELH Diagram.

Exhibit 4-14. Entity-life history diagram enhanced with trigger symbols.