When to Stop

Before beginning a significant project, write down what would make its purpose no longer justify its costs. Name who can stop it, what evidence they will see, and how a concern reaches them. If the only permitted outcome of review is continuation, it is not a stopping mechanism.

Revisit that account when the reasons change. A tool justified by one benefit can become a dependency maintained chiefly because ending it is now difficult. The difficulty is real. It is also a reason to inspect what has happened.

Reasons to stop and examine

Serious harm persists despite attempted mitigation. The affected people cannot obtain redress. Continuing requires concealment. The claimed benefit no longer has evidence behind it. The system is being used in a setting its evaluation did not cover, with no adequate protection for failure.

These are reasons for judgment, not automatic answers. Delaying treatment, access or a necessary repair can harm people too. Neither an irreversible choice nor uncertainty alone establishes that stopping is right.

Separate the costs of ending for users from the embarrassment of ending for its sponsors. Do not let the latter borrow the moral weight of the former.

End without disappearing

Where withdrawal is chosen, give appropriate notice, return data that can legitimately be returned, arrange alternatives where duties require them, and explain what will no longer be supported. A sudden shutdown can leave the burden with people who had no part in the decision.

The creator also needs an ending. Responsibility cannot mean being permanently available to every possible future demand. A clear handover can preserve accountability better than a tired promise to remain forever.

Stopping well includes the person who will inherit the absence.