When businesses invest in new tech they want to know it’s going to deliver. And so do the professionals who are developing the product. That’s why software testing is an integral part of the development of new digital technology.
Carried out throughout the project by technical people and end-users, testing is critical to the attainment of key milestones, including the project’s launch.
In this article we explore:
- why software testing is needed
- what is software testing?
- the different types of software testing
- common phases of testing
- standard testing procedures
Why is Software Testing Necessary?
American telephone network provider AT&T were well known for their provision of a reliable nationwide service. In January 1990, a single switch at one of their switching hubs suffered a minor mechanical malfunction which shut the centre down.
When the hub restarted itself and came back online, it sent out a signal that tripped out and reset switches in other centres’ operating systems. This created a devastating domino effect across the network that took out much of AT&T’s service.
The fault was traced back to a single line of faulty code from a complex software update that had been recently installed to speed up the network.
As a result, just one of AT&T’s customers, American Airlines, estimated that it lost 200,000 reservation calls. And AT&T lost the lustre of its reputation for providing a reliable service.
Had the code undergone a proper software test, the catastrophe could have been avoided. However, this story serves as a warning to other organisations: software testing is a critical part of any tech project.
What is Software Testing?
Software testing is a process, rather than a single activity, aimed at trying to find ways to make hardware or software fail. A systematic approach is used to run through common and not so common scenarios that software could be used for and carries out testing on each aspect of it.
This can include programme testing a particular software programme or performance testing elements of an entire product amongst a range of other kinds of testing.
Although there are lots of different ways to software test, the underlying goal is to establish whether a product delivers the intended outcome. Multiple different aspects can be tested: for example, if a point of sale system is upgraded to scan fingerprints, you would expect it to be able to scan fingerprints. But you might also want to ensure it meets other standards too. This can include checking for:
- Gaps between the solution and requirements by conducting usability, security or performance testing
- Defects that could occur if a programmer has made an error with some coding using programme testing
- Problems with interaction between different operating systems, software or programmes
Multiple software tests can take place throughout the project. These will usually be undertaken by programmers earlier on and, towards the end of the programme, also by users.
What Are the Different Types of Software Testing?
Testing is a broad church that encompasses many different approaches. Here are a few of the more common methods:
- A/B testing – this controlled experiment identifies whether a proposed change is more effective than a current method. Customers are sent to an existing version of a feature and a modified version of the same feature to see which is best at delivering the desired outcome.
- Performance testing – determines how a system performs under particular circumstances. This could include bombarding a system with lots of requests or placing it under an alternative strain to test its reliability and how much resource (like memory) it uses.
- Programme testing – identifies errors in code necessary to ensure the effectiveness of a programme that has been created.
- Usability testing – aims to ensure the user interface is easy to understand and use.
- Security testing – is essential for systems that contain confidential information. This testing is usually carried out by experts who evaluate whether unauthorised people can access the system and data. It also checks to see that authorised people can access the systems in the right way.
- Compatibility testing – software doesn’t exist in a bubble so it’s important to ensure it works well with other operating systems and applications and in different environments.
- Development testing – is performed by technical people, like a software developer, as software is constructed. The aim is to eliminate defects, improve the quality of the end result and enhance the efficiency of the overall software development process.
- Automated testing – removes the requirement for humans to manually test software or hardware by introducing robotics or programmes that can do it instead. It’s well suited for large projects that require the same areas to be tested repeatedly and reduces the risk of human error.
What is Standard Testing Procedure?
Every organisation will have their own software testing protocols. However, there is a common approach that follows a typical cycle to test software:
Step 1 – create a test plan to ensure every stage and requirement is covered off
Step 2 – prepare the procedures, test scenarios, cases, datasets and test scripts
Step 3 – execute the software testing using relevant testing tools: any issues are reported back to the development team
Step 4 – fix and re-test problems until all issues are resolved (or it’s agreed that certain fixes will be deferred) and the hardware or software is signed off and ready for release
Testing can sometimes take place in alpha and beta stages. Alpha testing is conducted by the creators of the tool or their client. Beta testing is when software is released to a small section of the public who will use the product to provide further real-world data for analysis and product improvement.
With sufficient testing completed, the product can be signed off ready for use, without any major glitches occurring.
Software testing is a core part of the development cycle of any digital product. Done right it can speed up the creation of upgrades and new offerings and deliver an end product that’s as close to perfect as possible.
To find out more about our industry-leading testing automation team, speak to an expert on +44 (0) 808 200 7375.