Databases Reference
In-Depth Information
RvsE
AvsE
N_aryRel
0,5
M:NRel
SCO
0
RR
CA
DA
MVA
IS_ARel
Figure 14.6
A Kiviat diagram.
14.2.6 Weighting
Weighting defines the relative importance of different quality factors in a
particular problem environment. It is impossible to say in absolute terms that
one data model is better than another, irrespective of context. Values can be
assessed only in the light of project goals and objectives. If the system under
development will be used as a basis for competing in the marketplace (e.g., a
product development system), then flexibility will be paramount. If the sys-
tem is used internally and the requirements are stable (e.g., a payroll system),
then flexibility will be less important. The concept of weightings helps to
define what is important and what is not important in the context of the
project.
Finding the best representation generally involves tradeoffs among
different qualities, and an understanding of project priorities is essential to
making those tradeoffs in a rational manner. Depending on users needs, the
importance of different qualities will vary greatly from one project to
another. Weightings provide the means to explicitly incorporate user priori-
ties into the evaluation process. An understanding of the relative importance
of different quality dimensions can highlight those areas where improvement
efforts will be most useful. The project team should come to a common
understanding of what is most important to the user as early as possible in
the modeling process. Ideally, the user sponsor should define the weightings
prior to any data modeling taking place. Analysts can then focus their efforts
on maximizing quality in the areas of highest value to the customer.
