Yay. I applaud this article. It explains that to build software well, one has to think for oneself, and not just follow some “guide” — that there is much more to building software. One must think about how testing is done, how one assesses coverage, and what technical practices could make things work better.

And my favorite point: that unit tests don’t test a system very well, especially if the system is a distributed real time system, which characterizes microservice and message oriented architectures: for such systems, the problems tend to be in the time-based interactions among the many parts, and unit tests and API contracts don’t cover that very well.