Software testing

People check computer programs to make sure they work.

They look for mistakes called bugs. These bugs can make a program slow.

Testing can happen in two ways. One way is to read the code. This is like proofreading a book.

The other way is to run the program. This shows if the program does its job. It checks if the results are right.

Testing can also look at how fast a program is. It can check if the program is safe. This helps make the software better.

Testing helps find many problems. But it cannot find every single bug. It is a way to make computers work well.

Software testing is a way to check computer programs.

It makes sure the software does what it should do. Testers look for mistakes called bugs. A bug is a flaw in the code. It can cause a program to act in a bad way. Bugs can also make a program run slowly.

There are two main ways to test. One way is called static testing. This is like proofreading a book. People review the code and the notes without running the program. The other way is dynamic testing. This happens when you actually run the software. You check if the real output matches what you expected.

Testers also use different views. In white-box testing, the tester can see the code.

They look at the inner parts of the program. In black-box testing, the tester does not see the code. They only see what the software does for the user. They provide inputs and check the results. Testing helps make software better and safer for everyone. It can even save a lot of money by finding errors early.

Software testing is a way to check if computer programs work correctly. It helps make sure software meets its goals and satisfies what users expect. Testing provides independent information about the quality of a program. It also helps people understand the risk that a program might fail.

While testing can check if a program works in certain situations, it cannot check every single possibility. This means testing cannot find every single bug. A bug is a defect in the code that causes an unwanted result.

There are different ways to perform these checks. One way is called static testing, which is like proofreading a document. In static testing, people review the code and the notes without running the program. Another way is dynamic testing, which happens when the program is actually running. Testers run the software to see if the real output matches the expected output.

There is also passive testing, where testers look at system logs without interacting with the software. This helps them find patterns or specific behaviors.

Testers often use different points of view to find errors. White-box testing means the tester can see the internal code of the program. They use their programming skills to choose inputs that test different paths through the code. This is similar to testing the individual nodes in an electrical circuit.

Black-box testing is different because the tester does not see the code. They only know what the software is supposed to do for the user. They provide inputs and check if the results match the requirements. Some people even use a hybrid method called grey-box testing.

People have been working to separate these tasks for a long time. In 1979, Glenford J. Myers introduced the idea of separating debugging from testing. Debugging is the work done to fix a bug once it is found. Before this, people did not always treat these as different activities.

Today, many companies outsource this work to places like China, India, the Philippines, and Pakistan to save on costs.

Testing is very important for the world economy. A study by NIST in 2002 found that software bugs cost the U.S. economy $59.5 billion every year. More than one-third of that cost could be avoided with better testing. Testing also helps with things like security and how fast a program runs. It ensures that software is easy to use and stays reliable.

By finding errors early, developers can build much better tools for everyone to use.

Software testing is the process of checking if software meets its intended objectives. It ensures that a program satisfies the expectations of its users or sponsors. This practice provides objective and independent information about software quality. It also helps stakeholders understand the risk of potential failure. Testing can confirm if software works correctly in specific scenarios. However, it cannot prove correctness for every possible scenario. It is impossible for testing to find every single bug.

To find problems, testers use an oracle to measure correctness. An oracle is a way to determine if a result is right or wrong. Common oracles include technical specifications, legal standards, or existing contracts. Testers might also compare a new product to past versions or similar products. They can use customer expectations or inferred purposes as guides. By using these oracles, testers can recognize when a program behaves incorrectly. This process helps answer if the software does what it is supposed to do.

Testing can be categorized by how the software is examined. Static testing involves reviewing code and documentation without running the program. This is similar to proofreading a text for errors. It includes activities like inspections, walkthroughs, or using tools to check code structure. Dynamic testing occurs when the software is actually executed. Testers run the code with specific test cases to see the actual output. They then compare this output to the expected result. Passive testing is another method where testers observe system logs and traces. They do not interact with the software but instead look for patterns.

Testers also use different perspectives called white-box and black-box testing. White-box testing, also called structural testing, examines the internal workings of the code. The tester uses programming skills to choose inputs that exercise specific paths through the program. This is like testing individual nodes in an electrical circuit. They may use techniques like API testing or fault injection to find errors. They also use code coverage to see how much of the program has been executed. A hybrid approach that uses both methods is called grey-box testing.

Black-box testing, or functional testing, ignores the internal code entirely. The tester only knows what the software is supposed to do for the user. They design test cases based on external descriptions like requirements or designs. This includes methods like boundary value analysis or fuzz testing. Another variation is component interface testing. This focuses on the data values passed between different parts of a system. Testers might check "message packets" to ensure data is valid before it moves to the next unit.

The history of this field changed significantly in 1979. Glenford J. Myers introduced the idea of separating testing from debugging. Debugging is the process of fixing a defect in the code. Before this, the two activities were often grouped together. Testing is used to find bugs, while debugging is used to fix them. A bug is a defect that causes an undesirable result. Not all defects lead to a failure. For example, a defect in "dead code" that never runs will not cause a failure. However, a defect might cause a failure later due to environmental changes.

Software testing is vital for the global economy. A 2002 study by NIST reported that bugs cost the U.S. economy $59.5 billion every year. The study suggested that more than one-third of this cost could be avoided with better testing. Because of these costs, many companies outsource testing to countries like China, India, Pakistan, or the Philippines. Testing also addresses non-functional requirements. These include scalability, security, performance, and maintainability. By using a "test pyramid," developers can organize tests into unit, integration, and end-to-end levels to maximize efficiency.