Micro-tests invite less coupling

I tend to advocate for very small SUTs when (unit) testing (a.k.a. micro-tests).

One of the reasons is that the fan-in factor from tests against large SUTs grows exponentially, thus affecting a very large amount of tests when redrawing design boundaries.

The smaller the SUT, the fewer tests its change is going to affect.

Interestingly, on some occasions, I’ve heard this as an argument for large SUTs. In my experience, the case is actually the opposite.