Building a Data Model - High-Performance Web Databases: Design, Development, and Deployment - page 217

This chapter has generalized the ANSI/X3/SPARC definition of a three-
schema architecture: ANSI/X3/SPARC defined the three types of schemas in
terms of how they should be supported within a given DBMS (e.g., a rela-
tional database product). The design principles of the three-schema frame-
work are incorporated to produce data models and an integrated business
conceptual schema that can be implemented through any (or, indeed, mul-
tiple) database technologies. Therefore, an architecture for DBMSs and
also a specification for effective database design practices are borrowed
from ANSI/X3/SPARC.

STEPS WITHIN LOGICAL DATA MODELING

The following list is a set of criteria for an optimal data model:

• Structural validity — Consistency with the way the business defines

and organizes information.

• Simplicity — Ease of understanding even by users or by nontechnical

professionals.

• Nonredundancy — Inclusion of no extraneous information and, in par-

ticular, representation of any one piece of information exactly once
(this may be a subcriterion of simplicity).

• Shareability — Not specific to any particular application or technology

and thereby usable by many.

• Extensibility — Ability to evolve to support new requirements with

minimal impact on the existing base.

• Integrity — Consistency with the way the business uses and manages

information values.

Satisfying these criteria requires more than assembling just any combina-
tion of boxes, arrows, and labels. Specifically, a set of rules and steps for
applying those rules should be followed to ensure an optimal data model.
Up to this point, this article has highlighted some of the rules, such as every
entity should have a primary identifier and primary identifiers are unique.
A sequence of steps for applying these rules is summarized as follows.

Building a data model involves examining
one activity or business function at a time. The information required by
each function is individually modeled, and the individual models are then
integrated into one composite model. The model or representation of infor-
mation required by one business function is known as a user view. Building
a user view begins with the following steps:

Building Skeletal User Views.

• Identifying major entities (significant objects of interest). (Step LDMI.)
• Determining relationships between entities. (Step LDM2.)

The process continues with the addition of
key detail information items and the most important business rules:

Adding Keys to User Views.