What is the right "unit" in unit testing and why it is not a class?
The talk argues that the right unit for testing is defined by scope and functional boundaries, not by classes, and suggests using heuristics to keep tests focused, stable, and valuable.
