Slide 9

CSCI 201 Guest Lecture: Data Modeling

Example: A bad data model

Good Problem

Peter's Pump Company (PPC) manufactures and sells electric powered water pumps for farm irrigation. PPC has six pump models, distinguished by their capacity in gallons per minute (GPM). Each pump also has a model number, an associated drawing (identified by drawing number), a price, and a quantity on hand (QOH). Whenever the QOH for a particular pump model goes below 10, Peter sends a work order to the manufacturing shop to build 20 more of that model. Peter wants a computer program to keep track of the relevant data and to automate the generation of work orders.

Bad Solution

Quick and Dirty Harry (QDH) comes up with a solution to the data modeling problem. He formulates PPC's data as a table:

GPM	Model	Drawing	Price	QOH
25	LX25	131467	$47	28
50	MX55	133519	$76	18
100	MX105	135221	$132	6
200	MX205	137733	$212	11
500	SX505	157393	$248	15
1000	XX1000	175418	$459	21

QDH thinks this is a really cool solution. It can be mapped directly to a 2D array (or could be mapped to a list of C structs). Data can be written to and read from disk easily, and in the array in memory, performance will be excellent for queries and updates. To add a new pump model, just increase the array's row dimension. To add a pump attribute, just increase the array's column dimension.

Email: Richard dot J dot Wagner at gmail dot com