DEV Community

Cover image for TDD Conference 2021 - On The Relationship Between Units Of Isolation And Test Coupling - Mario Cervera

TDD Conference 2021 - On The Relationship Between Units Of Isolation And Test Coupling - Mario Cervera

Maxi Contieri
Learn something new every day. - I am a senior software engineer working in industry, teaching and writing on software design, SOLID principles, DDD and TDD.
Originally published at ・2 min read

First International Test Driven Development took place on July 10th.

In this series, I will include every talk together with my notes and further reading.

Hopefully, a lot of readers will watch and rewatch the talks, as they are worth several reviews.

Let's continue...


Mario has been a software professional for more than a decade. He also conducted research in the field of Software Engineering, which allowed him to graduate as a PhD in Computer Science. As a strong advocate of software craftsmanship principles and values, he helps teams improve their technical practices so they can deliver higher-quality software at a sustainable pace.

TL;DR: Coupling leads to fragile tests.


My Personal notes

  • Fragility is the opposite to Robust when we talk about tests.
  • Fragile tests are tests that break when they should not break.
  • Tests are the specification of behaviour.
  • if we are not changing the observable behaviour of tests, and they break, the tests are preventing our refactorings.
  • We want to avoid, at all costs, fragile tests that fail when they shouldn't.
  • Coupling is the main cause of fragile tests.
  • Overspecified software makes more assumptions than necessary.
  • In a mocking approach, we need to change a lot of tests when we refactor.
  • if we make our attributes public to test them is a smell, we are testing accidental structure.
  • On Behavior-centric view, the true unit of isolation is not unit tests, but behaviour.
  • Refactoring is the key step in TDD.
  • Tests are code even more important than production code.

The outcome of this understanding is that a test-case per class approach fails to capture the ethos for TDD. Adding a new class is not the trigger for writing tests. The trigger is implementing a requirement.

Ian Cooper

Mario has also written an article on his talk.
Please rush to read it here.

Speaker Links

Please follow TDD Conference on:

Discussion (0)