How to Read a Tolerance

My father kept a set of gauge blocks on a shelf above his workbench. He never let me touch them as a child — not because they were dangerous, but because they were precise. A gauge block is a piece of hardened steel ground to an exact thickness, accurate to within a few millionths of a meter. You do not pick one up casually. You clean your hands, you breathe slowly, and you set it down on a surface that is itself verified flat. This is not ceremony. This is just what accuracy costs.

I thought of those blocks recently when a student came to me frustrated. She was twenty-two, training to maintain automated assembly systems, and she had failed the same calibration exercise three times. The equipment was not broken. Her measurements were not wildly wrong. They were simply inconsistent. Every attempt gave a slightly different number. She could not understand why, if she was doing everything correctly, the result kept changing.

I asked her: what does 'correctly' mean to you?

She looked at me the way students look when they think a question is too simple to be serious.

She said it meant following the procedure.

I told her that following a procedure is the beginning of correctness, not its completion. A procedure tells you what to do. It cannot tell you what to notice. Those are different skills, and one of them takes longer to develop.

Here is what I have learned from thirty years of working with machines, and what I try to teach now in a classroom with fluorescent lights that flicker slightly on cold mornings: every measurement has a tolerance. This is not a flaw in the system. Tolerance is the system's honest acknowledgment that perfection is a reference point, not an achievable state. The engineer's job — and I would argue the thoughtful person's job in any domain — is not to eliminate variation but to understand its range, name it, and decide whether it is acceptable.

My student's inconsistent readings came from something mundane. She was holding her breath during some measurements and not others. The small muscular tension in her shoulders changed how firmly she gripped the instrument. The instrument is sensitive enough to record that. She was introducing herself into the data without knowing it.

We are always introducing ourselves into the data. This is not a reason for despair. It is a reason for self-knowledge.

I have watched people treat self-improvement the way bad engineers treat a production line: they set an ideal specification, measure themselves against it, find the gap intolerable, and either force the process or abandon it. Neither response is useful. The useful response is to establish your baseline honestly, measure repeatedly, look at the spread of your results, and ask what is causing variation. Then you address the causes. Not the gap — the causes.

This sounds cold when I write it. I do not mean it coldly. My father, who left school at fifteen and spent four decades working in noise and heat and metal dust, understood it intuitively, without vocabulary like this. He knew that a person who cannot accurately describe their current condition cannot reliably improve it. He knew that wishful estimation is not the same as measurement. He knew these things because the machines told him, plainly, every day.

Machines are patient teachers because they do not flatter you. A dial gauge does not tell you that you almost got it right. It tells you the number.

After our conversation, my student went back to the calibration bench. She made a small change: before each measurement, she deliberately relaxed her shoulders and let her breath settle. Her results converged. Not to perfection — to an acceptable, consistent tolerance. She passed the exercise. More importantly, she could explain why she had failed before, in specific terms, which meant she would not fail the same way again.

That is the actual goal. Not the passing. The explanation.

I keep a single gauge block on my desk at the college, not to use, just to have. Students sometimes ask about it. I tell them it is a reminder that the standard is fixed and that we are the variable. Our task is not to resent that arrangement but to work carefully within it — measuring ourselves with something like the same patience we would extend to any system we are trying to understand.

Clean your hands. Breathe slowly. Set the instrument down on a verified surface.

Then read what it says.

Read in the app, leave a resonance, discover more stories.