One Concept Equals One Data Element: A Rule for Developing Data Elements - High-Performance Web Databases: Design, Development, and Deployment - page 81

and logical database design. As many data elements are eliminated and some
new data elements are created, new database designs can be contemplated.

Here, the key benefit obtained from ensuring that data elements comply
to the one-concept-equals-one-data-element rule is reducing redundancy.
Redundant data elements not only require more computer storage space,
they frequently lead to lower data quality when different sources are
updated at different times, allowing the data to become inconsistent.

The First Step

The problems caused by naming data elements before defining them
also must be addressed. People become attached to labels assigned to
objects and are reluctant to change those labels. Functional users, pro-
grammers, database administrators, and even data administrators usually
rely on the data element name as the unique identifier of the data element
(whether it is a full name or an abbreviated name). Therefore, the process
of naming data elements should not be taken lightly. A suggested sequence
of data element standardization involves the following steps:

• Developing the data element definition. This involves working with the

proposed data element custodian, steward, sponsor, or responsible
agent, ensuring that the data element definition includes only one
concept.

• Developing the valid values. This involves ensuring that they include

only one concept and are completely consistent with the definition.
Then the length, format, and edit criteria must be established.

• Developing the name from the definition. This involves selecting words

from the definition. All the words in the name should come from the
definition, but not all the words in the definition should be included in
the name. The urge to name the data element first should be resisted
and the definition from the name should be created. The processes are
not the same.

• Developing the remaining required attributes. These attributes differ,

depending on the organization.

Dependence on Names

An inverse relationship exists between the number of data element
attributes documented for a data element and the degree to which users
must depend on the data element name for meaning. As depicted in Exhibit
7, as more attributes are documented for a data element, the user needs to
depend on the name less and less to provide a complete understanding of
the data element. The significance of this relationship is that organizations
that are documenting (or intend to document) several attributes for their
data elements need not rely on a rigidly structured naming standard.