Slide 10

CSCI 201 Guest Lecture: Data Modeling

Example: A bad data model (cont.)

Peter Loved It

QDH coded his bad solution and installed the program at PPC. Peter loved it. It did everything exactly as he wanted. When the inventory got low for a pump, a work order was generated automatically, and the shelves were restocked.

Peter was encouraged with this success. He never really thought automation could work so well. There were just a couple of little things that could be improved...

Program Extension

Peter's pumps had several components that he bought from other suppliers. Sometimes his inventory of components ran out and that would halt production. Wouldn't it be nice if QDH's program could also track and flag pump component inventory? Some components were used in several pump models. Others were unique to a single model.

New Requirements Break the Bad Model

QDH thought about this new expanded problem. He could add a "components" column to the table and add a third dimension to the table to handle multiple components, or he could add a set of "component" columns to the table. This second option is less wasteful of memory than the 3D table option, but requires having enough component columns to cover the number of components in the pump with the most purchased components.

Another problem with extending the table model is that each component has its own attributes, so a fourth dimension to the table model would be necessary.

Email: Richard dot J dot Wagner at gmail dot com