The file itself is an inheritance: two hundred and forty-seven pages, organised by junction and then by recalibration scenario, passed across a desk in May of 2169 with a marginal layer its author had been adding to since the Persistent Error was commissioned. Its successor volume keeps the same order, the same subdivisions, the same marginal layer — chosen deliberately, so that whoever opens it next finds the same room arranged the same way. The sections are signed — HA. Not the teacher’s initials. Her own.
The covering letter
R —
The enclosed section is new. Not an update — new. At the far junction the array met a boundary response that nothing in your two hundred and forty-seven pages anticipates, and nothing in my fourteen either, and I have written the section the way you taught me to write the ones that matter: the profile met, the exact sequence of parameter moves, and then the thing underneath it, stated in the body and not tucked in the margin, because the next pair of hands should not have to deduce what I now know.
I would not have known how to begin a page with no precedent above it, except that beginning one is what the whole file is for. You will find the vocabulary is yours. The sentences are mine. I no longer apologise for the difference; I am told the difference is the work.
My brother’s pots continue to arrive on the supply relay in place of letters. I continue to read them with my hands. The mug from the observation deck is still on the ledge where you left it. Nobody drinks from it. You knew that when you left it there.
— HA
From the same working notes, routed through the ordinary channel
One further entry travelled the same season, two sentences in the routine maintenance log, filed to the drift-lab review in the ordinary way:
Calibration confidence drift, persistent across recalibration; baseline shift consistent with extended-mission instrument aging. Consult First Physicist at next drift-lab review for correlation against drift-lab register.
The stacks had come out of a full hardware refresh at Svalbard that spring. Fresh hardware is not supposed to age like that, nineteen days out. The entry stands as filed.
The reply
Read at ISNA Tromsø in 2174, and routed back on the relay in the teacher’s own hand:
You did the work.
Four words, against two hundred and forty-seven pages. The correspondence continued at the pace two careful people keep.