Every methodology has its own benefits and myths. The following sections will analyze the key benefits and most common myths of TDD.
TDD has its own advantages over regular development approaches. There are a number of benefits, which help in making a decision of using TDD over the traditional approach.
Automated testing: If you have seen a website code, you know that it's not easy to maintain and test all the scripts manually and keep them working. A tester may leave a few checks, but automated tests won't. Manual testing is error-prone and slow.
Lower cost of overall development: With TDD, the number of debugs is significantly decreased. You develop some code, run tests; if you fail, re-doing the development is significantly faster than debugging and fixing it. TDD aims at detecting defect and correcting them at an early stage, which costs much less than detecting and correcting at a later stage or post release.
Also, now debugging is much less frequent and a significant amount of time is saved. With the help of tools/test runners like Karma, JSTestDriver, and so on, running every JavaScript tests on browser is not needed, which saves significant time in validation and verification while the development goes on.
Increased productivity: Apart from time and financial benefits, TDD helps to increase productivity since the developer becomes more focused and tends to write quality code that passes and fulfills the requirement.
Clean, maintainable, and flexible code: Since tests are written first, production code is often very neat and simple. When a new piece of code is added, all the tests can be run at once to see if anything failed with the change.
Since we try to keep our tests atomic, and our methods also address a single goal, the code automatically becomes clean.
At the end of the application development, there will be thousands of test cases, which will guarantee that every piece of logic can be tested.
Improved quality and reduced bugs: Complex codes invite bugs. when developers change anything in neat and simple code, they tend to leave fewer or no bugs at all. They tend to focus on the purpose and write code to fulfill the requirement.
Keeps technical debt to minimum: This is one of the major benefits of TDD. Not writing unit tests and documentation is a big part, as this increases the technical debt for a software/project. Since TDD encourages you to write tests first, and if they are well written they act as documentation, you keep the technical debt for these to a minimum.
As Wikipedia says, "A technical debt can be defined as tasks to be performed before a unit can be called complete. If the debt is not repaid, interest also adds up and makes it harder to make changes at a later stage". More about technical debt can be found at https://en.wikipedia.org/wiki/Technical_debt.
Along with the benefits, TDD has some myths as well. Let's check few of them:
Complete code coverage: TDD forces the writing of tests first, developers write the minimum amount of code to pass the test, and almost 100% code coverage is achieved. But that does not guarantee that nothing has been missed and the code is bug free. Code coverage tools do not cover all the paths. There can be infinite possibilities in loops. Of course it's not possible and feasible to check all the paths, but a developer is supposed to take care of major and critical paths.
A developer is supposed to take care of business logic, flow, and process code most of the time. There is no need to test integration parts, setter-getter methods for properties, configurations, UI, and so on. Mocking and stubbing is to be used for integrations.
No need of debugging the code: Though test-first development makes one think that debugging is not needed, but it's not always true. You need to know the state of the system when a test failed. That will help you to correct and further develop the code.
No need for QA: TDD cannot always cover everything. QA plays a very important role in testing. UI defects, integration defects, are more likely to be caught by a QA. Even though developers are excellent, there are chances for errors. QA will try every kind of input and unexpected behavior that even a programmer would not cover with test cases. They will always try to crash the system with random inputs, and discover defects.
I can code faster without tests and can also validate for zero defect: While this may stand true for very small software and websites where the code is small and writing test cases may increase the overall time of development and delivery of the product. But for bigger products, it helps a lot to identify defects at a very early stage and provides a chance to correct them at a very low cost. As seen in the previous screenshots of the cost of fixing defects for phases and testing types, the cost of correcting a defect increases with time. Truly, whether TDD is required for a project or not depends on context.
TDD ensures good design and architecture: TDD encourages developers to write quality code, but it is not a replacement for good design practice and quality code. Will a team of developers be enough to ensure a stable and scalable architecture? Design should still be done by following the standard practices.
You need to write all tests first: Another myth says that you need to write all of the tests first and then the actual production code. Actually, generally an iterative approach is used. Write some tests first, then some code, run the tests, fix the code, run the tests, write more tests, and so on. With TDD, you always test parts of software and keep developing.
There are many myths, and covering all of them is not possible. The point is, TDD offers developers a better opportunity to deliver quality code. TDD helps organizations by delivering close to zero-defect products.