[Draiken]

What a weird claim.

PRs that are <50 lines are more often than not documentation, typos and small fixes in my experience.

Most real value add PRs end up bigger due to tests, documentation changes, and the actual change.

This for me is the typical claim that'll become an example of Goodhart's law when implemented. Random people that are not engineers will use this to claim "your PR should be 50 lines long" as some sort of absolute truth, when in reality the number of lines of code is absolutely irrelevant most of the time.

Context and details matter. Sometimes multiple small PRs are better. Sometimes a larger PR is better. It always depends. These absolutist claims that "this is better" is what leads to atrocious management practices that always end up getting summarized to a number, losing all of its meaning in the process.

Had we not we moved on from the lines of code metrics already decades ago? :facepalm: