All programs need to be thoroughly tested.

For example, in VB programming, if you only wanted to accept input values between 1 and 20 then you should display an error message if the user enters an invalid value.

A statement using VB syntax to produce this error message would be:-

MsgBox "Value should be between 1 and 20.  You entered " & InputValue & "Try Again!"

Testing

Testing strategy - A strategy to ensure that a final solution to a problem meets the original requirements of the user. 

Test plan - A document that details the scope, approach, resources, and schedule of intended test activities. Used to record the test data used and test outcomes. 

Types of Testing

Typical/standard data - Data that is valid and should be processed correctly by the program, used to check validation rules and to confirm accuracy by comparing results with previously calculated outcomes, i.e. testing for logical errors. 

Extreme/boundary data - Data that is valid but is at the extreme boundary of an acceptable range of values. Used to confirm data ranges have been correctly set. e.g. 1 or 5 for a film rating between 1- 5

Erroneous data - Data that would cause the program to fail if not validated and rejected. e.g. enter String "five" for a film rating with integer 1-5 input.

Test Data

Sort Algorithms - Duplicate values, accending order, decending order, negative numbers

Searching Algorithms - test with value in array, test with value not in the array

Algorithms - variety of data types (some expected others would be erroneous - see above) 

Types of Errors

Syntax Errors - Occur when the program code written by the programmer cannot be translated into source code.

Python example: PRINT(Hello World)

identifer should be lowercase i.e. print

Logical Error - This will compile and run but will not produce the expected outcome.

Example: a program intended to multiply two numbers contains the line:

answer = num1 + num2

Run-Time Errors - The program will again compile and run. The error occurs at run time and can be caused by impossible arithmetic calculations, overflows, or running out of memory. 

Example: divide by zero, accessing a memory address which does not exist.

Tools for Testing

Dry Run- Old fashioned paper and pencil. 

IDE tools - Watch and Trace, Cross Referencing, Setting Breakpoints. (see Unit 1.12 Software Engineering)

Dry Run

To dry run an algorithm you should list all inputs, processes and outputs in the columns of a  trace table. Then you can show each line number on a row (remember some lines will be excuted serveal times if iteration/looping is used).

Algorithm:

1. Declare Num1, Num2, Multiplier, Answer, Counter As    Integer

2. 

3.    Multipler = 2

4. Num1 = 1

5. Num2 = 4 

6.

7. Do While Multiplier < 4

8.     For Counter = Num1 to Num2 

9.         Answer = Counter * Multiplier

10.       Add_to_the_display: Counter & "Times" & Multiplier & " = " & Answer

11.   Next Counter

12    Multiplier = Mulitplier + 1

13.   Loop    

Trace Table:   

Test Data

Sort Algorithms - Duplicate values, accending order, decending order, negative numbers

Searching Algorithms - test with value in array, test with value not in the array

Algorithms - variety of data types (some expected others would be erroneous - see above)