Slide 2

CSCI 201 Guest Lecture: Data Modeling

What is a data model?

Human and computer thought and planning require abstraction from the real world. A model captures the essential aspects of reality that are relevant to the problem at hand.

As abstractions, models are necessarily simpler than the real world. It is their simplicity that makes them useful. Albert Einstein said that the best models are as simple as possible but no simpler. He was right.

Because models need to be simple, no single model can serve all purposes. However, a good model can grow, change, or adapt to new uses.

Computer Data Modeling Background

Computers were initially developed to do scientific and numerical computations (such as ballistics tables and orbital mechanics for space flight). Accordingly, computer languages of the period emphasized algebraic computation and numerical methods. Later on it was discovered that computers make excellent database managers for business applications. Database management system (DBMS) software was developed using various paradigms. The relational database paradigm has been the most successful of these, and recently the object oriented database (OODB) paradigm has seen a lot of development.

As computer applications have grown more ambitious, software development has grown more complex and difficult. People talk about the resulting "software crisis." Computer languages have evolved toward an object orientation in response. Object orientation enables software reuse. Modern object oriented languages include Smalltalk, C++, and Java.

Physics Modeling Example

A car is traveling on a highway at the rate (r) of a mile a minute (26.8 m/s). It's position on the road (p) at time t is described by the equation

p = r * t

This very simple model tells us exactly what we need to know about the car's location as a function of time. However, this model is unable to tell us the fuel remaining and can't account for the driver slowing due to traffic.

Email: Richard dot J dot Wagner at gmail dot com