We have the answers to your questions! - Don't miss our next open house about the data universe!

Test Driven Development” or the “Test First” policy

- Reading Time: 4 minutes
Test Driven Development" or the "Test First" policy

Software development is based on several stages characterized by the PDCA or Plan - Do - Check - Act principle. One of the most important steps on the Deming wheel is the verification (or testing) phase. To this end, there are a number of different methods for checking software quality and reliability. Test Driven Development is one of them. So what is Test Driven Development? What are the advantages and disadvantages of this method? And above all, how can it be implemented? DataScientest answers all your questions.

What is Test Driven Development?

TDD definition

Test Driven Development is a software development methodology based on a test-first policy. The idea is to guide software or application programming through a series of tests.

And this, even before starting to write the code.

In this way, as soon as bugs are encountered, developers correct them as they go along.

Regular testing means that any anomalies detected in the source code can be eliminated rapidly (almost in real time). In this respect, TDD perfectly meets the requirements of the Agile method, since it implies moving forward by iteration to improve the software progressively.

Good to know: This approach was initiated by the American developer Kent Beck. He simply reversed the process, starting by writing the test before writing the code. This allows him to improve code quality.

TDD and other test procedures

While Test Driven Development is a well-known DevOps method, it is far from being the only method for verifying software quality:

  • Acceptance Test Driven Development (or ATDD): a complementary method to TDD. In this case, developers focus on verifying the coherence of an application scenario with the targeted objectives. This approach focuses on functional application testing.
  • Behavior driven development (BDD): also known as behavior-driven programming, BDD focuses on the desired behavior, not on the correctness of the code.
  • Domain Driven Design (DDD): above all, this involves defining a common vision and language for all stakeholders.

All these methods complement each other. Together, they enable us to design better quality, scalable and reliable software, while reducing technical debt.

What are the advantages and disadvantages of TDD?


Today, Test Driven Development is recommended for the majority of programming processes. And for good reason: this testing method offers a multitude of benefits:

  • Anticipate software feasibility: by writing tests even before writing code, developers ensure that the very design of the software will enable them to achieve their objectives. Indeed, tests are more closely aligned with business needs. Above all, DevOps is able to cover more unit tests.
  • Save time: by writing tests upstream, you also avoid wasting time developing features that don’t meet your objectives.
  • Minimize errors: as development is test-driven, developers can detect bugs almost in real time. This limits post-deployment errors.
  • Simplify maintenance: as a corollary to the previous point, reducing bugs also saves time in terms of maintenance and over-servicing.


Despite all the advantages of Test Driven Development, we also need to be aware of its limitations.

On the one hand, this verification procedure implies an exhaustive understanding of the code. This implies a longer training period before starting to write the code.

On the other hand, the aim of test-driven deployment is primarily to verify the correctness of the code. This methodology does not focus on controlling the use of the software and its functionalities. This is why TDD generally needs to be complemented by other test procedures (such as BDT).

What are the 3 phases of test-driven development?

Test Driven Development is based on 3 fundamental laws:

  • “You must write a test that fails before you can write the corresponding production code”
  • “You must write only one assertion at a time, which causes the test to fail or fail to compile”.
  • “You must write the minimum amount of production code so that the assertion of the failing test is satisfied”.

Placed end-to-end in a single iteration, these 3 laws constitute a TDD nano-cycle. Traditionally, they are represented by a 3-phase cycle: red, green and refactor.

Phase Red

The first step in Test Driven Development is for the developer to describe a problem by writing a unit test. But since the code doesn’t yet exist, the test is bound to fail when executed. Hence the red symbol.

To carry out this test, the developer puts himself in the end-user’s shoes by describing the missing components essential to the code’s operation. But these are components that have not yet been implemented.

Phase Green

The aim of this phase is to pass the test. To achieve this, the developer writes a piece of code, completing the unit test written beforehand. It must pass. This is a research phase, in which the developer must find a solution to the problem posed.

But beware: during this phase of Test Driven Development, DevOps must not look for complex solutions. He must focus on the essentials, so that the production code turns green. He then verifies the code’s success by running further tests.

Phase refactor

Only once the solution has been found can the developer move on to the refactoring stage. He can then refactor the code, structure it, make it more readable and, if necessary, add to it.

However, he must not modify the code’s behavior. In other words, every feature must be retained. Here again, he must check that each test is validated each time a new feature is added.

Master TDD with DataScientest

The Test Driven Development approach fits in perfectly with the CI/CD (continuous integration and deployment) working method typical of DevOps. As a future developer, you need to master this approach in order to optimize code at every stage of development. To integrate it properly, it’s best to be trained in it. That’s why DataScientest offers you a training course that combines practical lessons with the best methodologies. Come and join us!

You are not available?

Leave us your e-mail, so that we can send you your new articles when they are published!
icon newsletter


Get monthly insider insights from experts directly in your mailbox