Testing Computer Programs
Understand the Different Types of Testing to Improve Development
Feb 13, 2009
Testing is an important and essential part of software development: further steps are explained in How to Develop a Computer Program. The purpose of testing is to ensure that the program works as expected, however this statement oversimplifies the possible complexities of thorough testing. Learning how to test well is a key skill for developers.
There are lots of different testing methods and there are many different types of test that can be undertaken. Good development includes choosing the appropriate tests for the particular program.
There are many reasons to understand testing and really get to grips with it. Firstly, testing ensure that programs work and secondly, understanding the full testing process will ensure that when programs go for further testing (by a test team or by users) then the program will pass those tests too. Knowing how others test, and why, will improve quality as programs are more likely to work first time at each stage.
Testing typically takes place in the following sequence:
Unit Test
A unit test is usually undertaken by the developer or coder and the purpose is to make sure that an individual piece of code works as expected. Unit testing is often performed at the same time as coding. Parts of the program are developed and tested and slowly built up to a working piece of code.
System Test
Once a number of pieces of code are ready and tested, then a system test can take place. The system test ensure that all the individually developed parts of a system work together.
If problems are found in system testing, then unit development is usually required and further unit and system testing will be necessary.
Functional Test
The system test usually focuses effort around at the points where the units meet and interact. Functional testing ensures that the system works ‘end to end’. Processes (or functions) are tested to ensure all the software and system works as expected.
Acceptance Test
The acceptance test is usually performed by the users of the system or the people who requested the development. The acceptance test is to ensure that the system performs as requested.
Wikipedia gives excellent explanations of two commonly used phrases in testing, ‘black box testing’ and ‘white box testing’. It is worth considering whether either is appropriate at different points of software development.
Understanding the different tests available and then selecting which tests to perform can help improve program quality. The sooner tests are completed successfully in software development, then the shorter the duration of the overall development cycle.
 |
