A developer deleted three hundred lines that no one was using.
The build broke in production.
The senior developer asked, “Was the code being used?”
“Yes, but by no one we knew.”
“Then who was the user?”
Beyond the Surface
Every codebase holds ghosts. Functions nobody remembers writing. Modules that show up in no roadmap, referenced by no ticket, understood by no one currently on the team. It is tempting to treat “unused” as a fact rather than a guess.
The developer in this koan did the responsible thing by every visible measure. Three hundred lines, no obvious caller, no complaints in code review. The deletion was clean. The build was not.
What broke was not the code. What broke was the assumption that visibility equals usage. Somewhere downstream, a job, a script, an integration partner, or a customer’s automation depended on behavior that had no name in anyone’s mental model of the system. The senior developer’s question, “then who was the user,” is not really about identifying a person. It points at something harder to see: systems accumulate users the way a river accumulates stones, quietly and without asking permission.
This is the deeper practice for anyone maintaining software at scale. Before removing something because no one seems to need it, ask what evidence supports that belief, and what evidence would look like if you were wrong. An empty pull request description is not proof of safety. A clean diff is not proof of correctness. Somewhere there may be a user you have never met, relying on a line you never noticed.