Databases Reference
In-Depth Information
Chapter 16
Making Data Models
Readable
David C. Hay
E
-
(
)
NTITY
RELATIONSHIP
MODELS
OR
SIMPLY
DATA
MODELS
ARE
POWER-
FUL
TOOLS
FOR
ANALYZING
AND
REPRESENTING
THE
STRUCTURE
OF
AN
.
ORGANIZATION
Properly used, they can reveal subtle relationships
between elements of a business. They can also form the basis for robust
and reliable database design.
Data models have gotten a bad reputation in recent years, however, as
many people have found them to be more trouble to produce and less ben-
eficial than promised. Discouraged, people have gone on to other
approaches to developing systems — often abandoning modeling alto-
gether, in favor of simply starting with system design.
Requirements analysis remains important, however, if systems are ulti-
mately to do something useful for the company. And modeling — especially
data modeling — is an essential component of requirements analysis.
It is important, therefore, to try to understand why data models have
been getting such a bad rap. This author believes, along with Tufte, that the
fault lies in the way most people design the models, not in the underlying
complexity of what is being represented.
GOALS OF ENTITY-RELATIONSHIP MODELS
An entity-relationship model has two primary objectives:
• First it represents the analyst's public understanding of an enterprise,
so that the ultimate consumer of a prospective computer system can be
sure that the analyst got it right. Does the analyst really know the busi-
ness? A properly drawn data model can provide a vivid answer to that
question. To the user it addresses the question: “Is this what you want?”
• Second, it represents the fundamental architecture of an enterprise's
data, and as such is an ideal starting place for database design. To be
sure, the final structure of a database may differ from the data model
for good and valid reasons, but the closer it remains to the original
Search WWH ::
Custom Search