3.9 Techniques for Testing

3.9 Techniques for Testing

1. Test Harness

   • Test harness allows some types of testing to be automated - it is sometimes know as automated test framework
   • A test harness consists of software and test data which run code under different conditions - monitors the actual output against expected output
   • Removes the need for a programmer to do repetitive tasks, therefore saving time and money

2. Volume Tetsing

   • Designed to show how the system will perform with different volumes of data - both large and small amounts
   • Designed to establish the limits of the software under a large volume of data - to ensure it meets the specification, to check a system can cope with the projected capacity of data that is expected

3. Scalability Testing

   • Performance testing focused on making sure the system is able to cope with increases in workload - checking it will function correctly when the system is scaled up
   • A series of tests are carried out as the system is scaled up - until the performance levels start to drop at which point test results are analysed, the analysis will identify bottlenecks that are preventing the system from being scaled up

4. Prototyping

   • Not strictly speaking a testing term - but it is a useful technique for gathering feedback about the user interface from clients/users
   • Involves producing working but limited versions of a system at different stages during its development - allowing for feedback to be gathered throughout the development process, required changes are made and the client/user is shown again, this is an iterative development methodology

5. Multi-Platform Testing

   • Sometimes known as Portability testing
   • Many systems need to operate on a variety of platforms - e.g. systems software: windows, Apple OS, Linux, e.g. hardware: pc, tablet, smartphone
   • Aims to demonstrate that software can be transferred across (ported) to specified hardware and software platforms

6. Use of Simulated Environments

   • Not always feasible to test the system in every environment it could possibly meet so simulated environments are useful - if system needs to run on the client's WAN, it is difficult and time-consuming to install and test, difficult to control the exact conditions that the software might meet
   • Using simulated environments means that the test conditions can be repeated over and over again - using various configurations as fair comparisons
   • Used to see how the system behaves when it is subjected to a series of controlled inputs
   • Any errors the system produces can be dealt with by developers