Memorandum on Unused Variable Suspicion
An essay in three sections on the institutional logic of hesitating before a deletion that appears obviously safe.
I.
The variable is unused. The linter confirms this. The code compiled before it appeared and will compile after it is gone. There is no argument for keeping it. The delete key is available.
The developer hesitates.
This is the moment under examination. Not the deletion, which will almost certainly occur. Not the linter warning, which is accurate. The hesitation is the event of interest because it arrives before any evidence that something is wrong and it arrives every time.
Its uselessness seems almost too obvious, which is the first indication that something is wrong with the variable.
II.
The workaround has a logic to it. Unused variables that are genuinely unused do not usually attract suspicion. They are deleted on sight. The variables that produce hesitation are the ones that look unused but carry context the reader does not yet have: a framework that reads it by reflection, a test that will restore it, a function that was removed last week and will return next week. The developer has been wrong about these before.
The workaround reduces immediate risk. It does this by adding one extra decision to a moment that previously required none. Other people begin anticipating that decision and quietly adjusting their own behavior. They stop naming variables in ways that invite this outcome. The variable name acquires a small secondary meaning: a variable that looks unused and is named this way is one that should be checked before deletion.
The workaround produces a second check designed to confirm the first one worked.
III.
Eventually the ordinary task has a parallel administrative system built around unused variable suspicion. Deleting a variable now takes longer than it did before the suspicion became standard procedure. Everyone knows this and no one names it, because naming it would require acknowledging that the procedural overhead was invented from accumulated caution rather than from any policy.
Under the governing rule, unused variable suspicion becomes valid procedure after two repetitions without objection. It has exceeded this threshold. No objection is on record. The behavior continues.
The ordinary task feels improperly performed unless unused variable suspicion happens first. This feeling is not in the documentation. It did not come from anywhere. It is simply present.