Marginalia

How you got here

You read a comment. A comment is the part of a document addressed to nobody — present in the source, absent from the page, written by someone who assumed it would be read only by a person already inside.

You were not inside. You looked anyway.

The document

The ARGENT materials group ran a sequence of thermal cycling tests through 2001 and reported the results in March 2002. The report was careful, internally consistent, and wrong. It attributed a drift in the measured expansion coefficient to the sample. The drift was in the fixture.

The error stood for eleven months. It was corrected in February 2003 when a new member of the group, reading back through the bound laboratory notebooks because nobody had told her not to, found this in the margin of the entry for 14 August 2001, in pencil, in a hand nobody has since identified:

fixture? check w/ blank. clamp gets warm.

Three words and a question mark. The blank test takes forty minutes. It had never been run, because the formal procedure did not call for it, and the person who wrote the note did not raise it, and there is no record that they ever mentioned it to anyone.

The part we published

The correction was routine and is not why this document exists. What the Directorate examined afterwards was the eleven months, and specifically why a correct observation, written down in the correct place at the correct time, took eleven months to reach anyone.

The answer was not secrecy or incompetence. It was that the notebook margin is not an output. Nothing in the institution’s process reads a margin. The margin is where you put a thought that is not yet a claim, and the institution had no route by which a thought that is not yet a claim could become one, short of the person carrying it themselves.

Most people, most of the time, do not carry it themselves. They write it in the margin and go to lunch. This is not a character flaw. It is what a margin is for.

What changed

Two things, both small.

Bound notebooks at all sites are now scanned in full at the close of a programme, margins included, and the scans are indexed as text. The index is searched at the start of any review that reopens a closed result.

And the review template gained one line, which reads: what did someone write down and not say out loud?

It is a strange question to put on a form. It has been answered usefully fourteen times.

We had built an institution that was extremely good at hearing claims and had no ear at all for doubts. The pencil note was not a claim. It was a person being uncertain in the only place where uncertainty was allowed, which was a place we had agreed not to read.

— ██████ ██████, Directorate VI

Why it is unlisted

One remains. It is the quietest of the four, and it is written in the space this format reserves for nothing at all.

00 00 00 00 00 00 53 45 4e 2d 32 30 2d 30 30 30 00 00 00 00