Types of Software Testing
What Is Software Testing?
Software testing is a fast-paced, growing field. Many times there can be a misunderstanding between testers, devs, and management about testing terminology. Having clear, shared definitions of testing is important because of this.
This article will go over some of the common types of software testing. It's important to keep in mind that in practice, it can be easy for two (or more) types of testing to overlap. Academically these are all very distinct terms, but pragmatically they can begin to merge.
I also will not talk specifically about "automation" as a testing type, since most of these types of testing can be accomplished either manually or automatically with code.
Common Types of Software Testing
- Functional Testing
- Integration Testing
- Black box Testing
- White box Testing
- Regression Testing
- Smoke Testing
- Exploratory Testing
- User Acceptance Testing
Note: This is not a complete list of the types of software testing.
What Is Functional Testing?
Functional testing is a sub-group of other testing types. Functional testing is any black-box (not code side) testing that makes sure the software completes the appropriate functions. This is the type of testing that makes sure the requirements are met.
Types of functional testing include:
- Sanity testing
- Smoke testing
- Regression testing
- Integration testing
- Usability testing
What Is Integration Testing?
Integration testing is incredibly important for software that works on a server-client model, or as a distributed system. This means applications that do not live solely on the user's computer and applications that may have multiple different servers on the back end.
An architecture example could be:
- The main server that is running a web application's dashboard.
- A separate server that solely holds the database.
- A separate server that runs reports or algorithms for the system.
- An additional piece of software that allows for this application to connect with a third-party application.
Integration testing for the scenario outlined above would include making sure each server is able to talk to the appropriate other servers, that the correct information is passed from one service to the other, etc.
Integration testing can also occur at a smaller level, once developers have completed microservices or small bits of code that need to interact with other parts.
Black Box Testing
Black box testing is a broad term for any testing where the code is not looked at.
The idea is that the product is like a black box. The user does not care what happens in the black box so long as they get the correct output from their input.
Most functional, manual test cases are black box testing.
White Box Testing
White box testing focuses on the internal properties of the application, rather than the functionality of using the product. One of the main benefits of white box testing is that it discovers errors before they appear at the level of the user. This is especially useful for hidden errors that may not cause visible defects now, but cause defects with further changes to the code.
Unit testing is the most common type of white box testing.
Regression testing's focus is on making sure that no new features, architecture changes, or recent defect fixes have accidentally created new defects. Many times a code change can have a cascading effect and create defects in an area that may be missed when that fix is tested.
Regression testing is very useful if many changes have been made to a product in multiple areas. This can be the most time-consuming area of testing though. Worse case scenario the entire product has to be tested for all requirements.
A highly valuable skill is the ability to balance the correct amount and types of testing for a regression test.
Regression testing is usually utilized directly before a product is released.
Smoke testing, also known as sanity testing, is a high-level form of testing. Smoke testing completes the happy path testing for the most important functions of the product while looking for any critical defects at that level.
This helps to ensure that an obvious critical bug isn't discovered halfway through the testing when a quick click through would have shown it.
Exploratory testing is when a tester is given an area or functionality of a product to explore. They do not follow detailed test cases but instead may use them as a loosely defined script. Exploratory testing does tend to focus on a lot of negative testing.
This type of testing tends to focus on the question "I wonder what would happen if I?" Testers look for things that aren't covered in the requirements. Such as adding a product and removing the product from an online shopping cart five times, then trying to purchase something else.
User Acceptance Testing
User acceptance testing (UAT) is a final step in the testing process. This is when the final user for a product uses the product. The product should be stable and ready to go to production at this point.
Once the user "accepts" the product, it ready to go to production.
This is more common practice with projects that have a single client.