Software testing involves evaluating and verifying a software product’s functionality. Basically, it checks whether the software product matches anticipated requirements and makes sure it is defect-free. It can be said that testing enhances the quality of the product by preventing bugs, reducing development costs, and reducing performance issues.  

You can test the software in many different ways. Some types of testing are conducted by software developers and some by specialized quality assurance staff. Here are a few different kinds of software testing, along with a brief description of each.

Software testing is governed by seven principles:  

• Absence of errors fallacy: Even if the software is 99% bug-free, it is unusable if it does not conform to the user’s requirements. Software needs to be bug-free 99% of the time, and it must also meet all customer requirements.
• Testing shows the presence of errors: Testing can verify the presence of defects in software, but it cannot guarantee that the software is defect-free. Testing can minimize the number of defects, but it can’t remove them all. 
• Exhaustive testing is not possible: The software cannot be tested exhaustively, which means all possible test cases cannot be covered. Testing can only be done with a select few test cases, and it’s assumed that the software will produce the right output in all cases. Taking the software through every test case will cost more, take more effort, etc., which makes it impractical.
• Defect clustering: The majority of defects are typically found in a small number of modules in a project. According to the Pareto Principle, 80% of software defects arise from 20% of modules.
• Pesticide Paradox: It is impossible to find new bugs by re-running the same test cases over and over again. Thus, updating or adding new test cases is necessary in order to find new bugs.
• Early testing: Early testing is crucial to finding the defect in the software. In the early stages of SDLC, defects will be detected more easily and at a lower cost. Software testing should start at the initial phase of software development, which is the requirement analysis phase.
• Testing is context-dependent: The testing approach varies depending on the software development context. Software needs to be tested differently depending on its type. For instance, an ed-tech site is tested differently than an Android app.

The dictionary definition of regression is the act of going back to a previous place or state. In software, regression implies that a feature that used to work suddenly stopped working after a developer added a new code or functionality to the software.

Regression problems are pervasive in the software industry, as new features are getting added all the time. Developers don’t build these features in isolation, separate from the existing code. Instead, the new code interacts with the legacy code and modifies it in various ways, introducing side effects, whether intended or not.

As a result, there is always a chance that introducing new changes may negatively impact a working feature. It’s important to keep in mind that even a small change has the potential to cause regression.

Regression testing helps ensure that the new code or modifications to the existing code don’t break the present behaviour. It allows the tester to verify that the new code plays well with the legacy code.

Imagine a tourist in a foreign city. There are two ways in which they can explore the city.

• Follow a map, itinerary, or a list of places they should visit
• Explore randomly, following the streets as they lead them to new places

With the first approach, the tourist follows a predetermined plan and executes it. Though they may visit famous spots, they might miss out on hidden, more exciting places in the city. With the second approach, the tourist wanders around the city and might encounter strange and exotic places that the itinerary would have missed.

Both approaches have their pros and cons.

A tester is similar to a tourist when they are testing software. They can follow a strict set of test cases and test the software according to them, with the provided inputs and outputs, or they can explore the software.

When a tester doesn’t use the test scripts or a predefined test plan and randomly tests the software, it is called exploratory testing. As the name suggests, the tester is exploring the software as an end-user would. It’s a form of black-box testing.

In exploratory testing, the tester interacts with the software in whatever manner they want and follows the software’s instructions to navigate various paths and functionality. They don’t have a strict plan at hand.

Exploratory testing primarily focuses on behavioural testing. It is effective for getting familiar with new software features. It also provides a high-level overview of the system that helps evaluate and quickly learn the software.

Though it seems random, exploratory testing can be powerful in an experienced and skilled tester’s hands. As it’s performed without any preconceived notions of what software should and shouldn’t do, it allows greater flexibility for the tester to discover hidden paths and problems along those paths.

End to End testing is the process of testing a software system from start to finish. The tester tests the software just like an end-user would. For example, to test a desktop software, the tester would install the software as the user would, open it, use the application as intended, and verify the behavior. Same for a web application.

There is an important difference between end-to-end testing vs. other forms of testing that are more isolated, such as unit testing. In end-to-end testing, the software is tested along with all its dependencies and integrations, such as databases, networks, file systems, and other external services.

Unit testing is the process of testing a single unit of code in an isolated manner. The unit of code can be a method, a class, or a module. Unit testing aims to focus on the smallest building blocks of code to get confidence to combine them later to produce fully functioning software.

A unit test invokes the code and verifies the result with the expected result. If the expected and actual outcomes match, then the unit test passes. Otherwise, it fails.

A good unit test has the following characteristics:

1. It should test a single piece of functionality.
2. It is fully automated and repeatable. 
3. It should run quickly and provide immediate feedback.
4. It should be isolated and shouldn’t interact with external dependencies such as network, database, or file system unless needed. You can use the mocking technique to simulate the external dependencies and isolate the code under test.

API stands for Application Programming Interface. It is a means of communication between two software components. An API abstracts the internal workings and complexity of a software program and allows the user of that API to solely focus on the inputs and outputs required to use it. 

When building software, developers rarely write software from scratch and make use of other third-party libraries. An API allows two software components to talk to each other by providing an interface that they can understand.

Another use of an API is to provide data required by an application. Let’s say you are building a weather application that displays the temperature. Instead of building the technology to collect the temperature yourself, you’d access the API provided by the meteorological institute.

A test environment consists of a server/computer on which a tester runs their tests. It is different from a development machine and tries to represent the actual hardware on which the software will run; once it’s in production.

Whenever a new build of the software is released, the tester updates the test environment with the latest build and runs the regression tests suite. Once it passes, the tester moves on to testing new functionality. 

When software is being tested, the code coverage measures how much of the program’s source code is covered by the test plan. Code coverage testing runs in parallel with actual product testing. Using the code coverage tool, you can monitor the execution of statements in your source code. A complete report of the pending statements, along with the coverage percentage, is provided at the end of the final testing.

Among the different types of test coverage techniques are:

• Statement/Block Coverage: Measures how many statements in the source code have been successfully executed and tested.
• Decision/Branch Coverage: This metric measures how many decision control structures were successfully executed and tested.
• Path Coverage: This ensures that the tests are conducted on every possible route through a section of the code.
• Function coverage: It measures how many functions in the source code have been executed and tested at least once.

• Black-box testing in software testing: In black-box testing, the system is tested only in terms of its external behaviour; it does not consider how the software functions on the inside. This is the only limitation of the black-box test. It is used in Acceptance Testing and System Testing.
• White-box testing in software testing: A white-box test is a method of testing a program that takes into account its internal workings as part of its review. It is used in integration testing and unit testing.
• Grey-box testing in software testing: A Gray Box Testing technique can be characterized as a combination of a black box as well as a white box testing technique used in the software testing process. Using this technique, you can test a software product or application with a partial understanding of its internal structure.

It is extremely beneficial to use automation testing when using the agile model in software testing. It helps in achieving maximum test coverage in a lesser time of the sprint.

• Test Case: Test Cases are a series of actions executed during software development to verify a particular feature or function. A test case consists of test steps, test data, preconditions, and postconditions designed to verify a specific requirement.
• Test Scenario: Usually, a test scenario consists of a set of test cases covering the end-to-end functionality of a software application. A test scenario provides a high-level overview of what needs to be tested.
• Test Scripts: When it comes to software testing, a test script refers to the set of instructions that will be followed in order to verify that the system under test performs as expected. The document outlines each step to be taken and the expected results. 

A software bug is an error in the software that produces wrong results. A software tester tests the software to find bugs in it.

There are many causes for the bugs—for example, poor design, sloppy programming, lack of version control, or miscommunication. Throughout development, developers introduce hundreds or thousands of bugs in the system. The goal of the tester is to uncover those bugs.

You can find a bug in many different ways, regardless of your role. When building the software, the software developer might notice the bug in another module, written by another developer or by themselves. The tester actively tries to find the bugs as part of a routine testing process. Finally, the users could see the bugs when the software is in production.

All bugs, no matter how they are found, are recorded into a bug-tracking system. A triage team triages the bugs and assigns a priority to the bug, and assigns the bug to a software developer to fix it. Once the developer resolves the problem, they check in the code and mark that bug as ready for testing. Once a bug is ready for testing, it goes to the tester, who tests the software to verify if it’s indeed fixed. If it is, then it’s closed. If not, they assign it to the same developer with a description of the exact steps to reproduce the bug. Some examples of popular bug-tracking systems include BugZilla, FogBugz, etc.

Trivia:

The first software bug was discovered by Admiral Grace Hopper, on September 9, 1947. After they opened a malfunctioning piece of hardware, they found an insect stuck in the relay.