The feature no one noticed for two years

There is a small piece of behavior in the system — a rule about how a certain kind of conflict resolves itself when two practitioners edit the same surface within a few seconds of each other. It was written carefully. It was tested for a week before it shipped. It has run, without modification, for a little over two years.

No one has ever mentioned it. No support ticket has referenced it. No retrospective has cited it. By the standards of the metrics we were taught to care about, it does not exist.

The first year, this bothered me. The work had been good and the work had been invisible, and the lesson the industry teaches is that invisible work is wasted work. I considered, more than once, redesigning it to be more legible — adding a small notification, a subtle animation, anything that would let a practitioner notice that something thoughtful had happened on their behalf.

I am glad I did not. The behavior works because it does not announce itself. The moment it asks for credit, it becomes a feature the practitioner has to think about, which is the opposite of what it was designed to do. Its silence is not a failure of communication. Its silence is the form the value takes.

The moment it asks for credit, it becomes a feature the practitioner has to think about — which is the opposite of what it was designed to do.

This is SHIBUI as a working principle, not an aesthetic theory. The bowl that sits on the shelf and is reached for every morning. The function in the codebase that has not changed in a decade because the requirements never asked it to. The instructor whose corrections you only understand were corrections years after they were made.

Most of what is durable in a long practice will not be noticed in the quarter it ships. Some of it will not be noticed in the decade it ships. The practitioner who needs the work to be noticed will eventually stop building work like this, and the practice will become shallower for it.

The question worth asking, before any non-trivial decision, is not whether the work will be impressive. It is whether the work will still be working — quietly, without thanks — in ten years.