Databases Reference
In-Depth Information
Chapter 18
Avoiding Pitfalls in
Data Modeling
Steven Cheung
D
,
ATA
MODELS
ARE
USED
TO
IMPROVE
THE
LOGICAL
DESIGN
ORGANIZA-
,
'
.
TION
When develop-
ing such devices, however, the data modeler can encounter many obsta-
cles. This article suggests ways to create a high-quality model while
avoiding the stumbling blocks.
AND
IMPLEMENTATION
OF
A
CORPORATION
S
DATA
CRITICAL CONSIDERATIONS TO IMPROVE DATA MODEL QUALITY
To help a data modeler produce a high-quality model, several critical
areas must be addressed. These areas — semantics, technique, complete-
ness, and accuracy — and their considerations are discussed more thor-
oughly in the following sections.
Semantics
Semantics deals primarily with the proper identification and naming of
data modeling objects. A high-quality data model requires precision in the
choice of such names for objects as entity, attributes, and relationships.
A precise name conveys its meaning unambiguously, and difficulty in arriv-
ing at such a name might suggest a lack of understanding of the object's
definition. Semantics is particularly important in differentiating between a
concrete object and an abstract object. Throughout this article, examples
of proper semantics are discussed.
Technique
An important element contributing to the data modeling process is how
effectively the model reflects the business rules. Examples of a particular
business situation's being most effectively modeled by a particular choice
of modeling technique include the handling of the time dimension in data
modeling, the use of subtypes, and recursive relationships.
Search WWH ::
Custom Search