Databases Reference
In-Depth Information
8.4.3.
The blank page effect
To counter the blank page effect, the data analyst should
be given a pre-built data model, ready for use, serving him as
a base for his modeling work. Moreover, the blank page
effect is often a negative factor when it is necessary to start a
negotiation with multiple actors on data modeling that must
be shared. If this negotiation starts as a blank page, without
any prior proposal recognized as being acceptable or
desirable, then the negotiation can be long and delicate.
Also, it is best to avoid the temptation of re-engineering an
existing database in order to obtain a first data model.
This approach is often destined for failure as what is
already in place contains too many errors and useless
complexities collected over time.
The objective of semantic modeling is to offload, once and
for all, the technical and legacy constraints in order to re-
appropriate business fundamentals.
To counter the blank page effect, it is necessary to use
data models that are ready to use, called “patterns”. Unlike
other technical patterns already known to software
development, such as “Factory”, “Facade” or “Singleton”,
these are semantic modeling patterns.
8.5. Ready-to-use semantic models
Before reviewing possible reusable data model sources, it
is necessary to specify that we are looking for a lot more than
data models described in technical terms. We need true
semantic models expressed with standard UML notation,
without data duplication. These models must not be
cluttered with join tables which involve a technical
complication that we do not need.
