Slide 6

CSCI 201 Guest Lecture: Data Modeling

Why data modeling?

We have already seen where, in the robotic navigation problem, the problem nearly solves itself with the proper selection of a model. Proper modeling of the problem domain is one of the most important steps in software development.

The software development process usually involves iterating through a series of steps. The process usually goes something like this:

Define Requirements
Create a Data Model
Code the Prototype
Test the Prototype with Actual Users
Redefine Requirements
Modify the Data Model
Modify the Code
Retest

Problem Domains are Complex

The previous examples had simple problem domains. The problem domains you encounter in practice can become extremely complicated. Modeling allows the use of simplifying abstractions while letting you determine if the model captures the essential features of the problem.

Requirements Change

Requirements always change. Assume from the beginning that the requirements will change. New requirments are added. Sometimes requirements are dropped. A good model will let you make these changes.

Problem Domains Change

In the robot navigation problem, the cars are parked. Suppose we add cars that are moving? In this case, the model needs to be modified to add velocity information to the cars.

In the school scheduling problem, suppose the computer science department merges with the philosophy department (just suppose). Can the model handle the domain change?

Email: Richard dot J dot Wagner at gmail dot com