COMPLETING THE LOGICAL DESIGN - Database Design and Development: An Essential Guide for IT Professionals - page 331

Real-world
Information

FEASIBILITY
STUDY

REQUIREMENTS
DEFINITION

PLANNING

Plan Document Study Document Data Groups

LOGICAL
DESIGN

Conventional
Data Model

Semantic
Data Model

Figure 11-4 DDLC steps up to logical design.

designing the relational data model, you would have created it with the data nor-
malization method. In any case, you now have a relational data model. Figure 11-4
recapitulates these steps.
Having gone through these distinct steps, how are you sure that the output of the
logical design phase truly and completely represents the real-world information?
Has something been missed along the way? Has something been misrepresented in
the transition from step to step? Especially if the database project is large with
multiple analysts working in each step of DDLC, there are more chances of things
going wrong. So, this is a good stage in DDLC to pause and ensure the complete-
ness and correctness of the design. Adjustments can be made with comparative ease
at this point rather than after defining the files and records in the physical design
phase.
Go back and review the outputs from all the steps leading up to this point and
adjust the logical design accordingly. Pay attention to the following critical factors
in your review:

• Real-world information requirements could create variations in the way busi-
ness objects and the relationships among these are perceived in your organi-
zation. Ensure that your semantic data model represents all the nuances of the
information requirements in your particular environment.
• Semantic data modeling techniques have all the necessary components to
represent every aspect of real-world information. Become familiar with the
features and purposes of the data model components and employ these
suitably.
• If the scope of the information requirements for your organization is large and
complex, your database project team will be dividing up the data modeling task.
The team will be creating partial data models. Ensure that nothing gets lost