Supporting Your Code, README vs Wiki and Test Coverage

Misconceptions

Test coverage percentages often mislead developers into believing that a high number equates to quality code. Allen Underwood and Joe Zack discuss how setting unrealistic test coverage goals can lead to gaming the system, where developers write superficial tests just to meet targets 1. Joe argues that a target percentage is not inherently bad, but it becomes problematic when it's set too high, as it encourages shortcuts rather than meaningful testing 2.

You can completely game that. And you're not, you know, really getting the goal that you want.

--- Joe Zack

Allen adds that this approach is akin to enforcing bad rules, which ultimately demotivates teams and leads to ineffective testing practices 2.

   

Realities & Practices

The misconception that 100% test coverage guarantees bug-free code is debunked by Joe and Allen. They illustrate this with a simple function example, showing that even fully covered code can fail under untested conditions, like dividing by zero 3. Allen emphasizes that test coverage should focus on critical business rules rather than superficial metrics, as thorough testing of essential code is more valuable than blanket coverage 3.

100% coverage doesn't mean that you thought about all the cases.

--- Allen Underwood

Additionally, they touch on best practices for robust testing, highlighting the importance of meaningful tests that ensure confidence in code changes rather than just achieving high coverage numbers 4.