Instead, I logged the note. I wrote down the date, the fact that it had been placed in my mailbox by hand, and the sentence itself, in quotation marks, word for word. I put the note itself in the folder with the certified-mail receipt and the return receipt card, because it belonged there now, as part of the record. It was not a reply to the letter in any substantive sense. It was a response to being told no, and it deserved to be filed as such.
That evening, I sat in my kitchen and thought about the phrase "like adults." Adults, in my experience, kept records. Adults signed for letters. Adults understood that permission, once withdrawn in writing, was not restored by a request for a chat. Adults did not use a pool at 12:40 a.m. and then ask to talk about it like adults. I did not write any of this down. I did not need to. The note was in the folder. The log entry was beside it. The silence of the four days before it was noted by its absence.
Sandra's note did not change the letter. It did not reopen the question. It simply confirmed what the signed receipt had already suggested: she had read it, she had understood it, and she did not accept it. That was useful to know, and I filed it accordingly.
In the days that followed I kept my routine deliberately ordinary: work, home, the kitchen window, the notebook. I did not patrol the fence or watch for Sandra's movements, because watching would have turned the record into surveillance in my own mind, and I wanted it to remain what it was, a log of facts noticed in the course of an ordinary life. If Sandra knocked, I would answer politely. If she did not, I would not chase her. The note had asked for a talk, and my answer was the absence of a talk, recorded in the file by the note itself. Silence, filed properly, is also a response, and it has the advantage of not giving the other party new sentences to rearrange.
Priya Nair came to my house a few days after the mailbox note, and she said the sentence that changed how I kept the record.
Priya is my friend and a fellow accountant, which means she understands, without explanation, why a person would keep receipts for twelve years and file them in order. She had listened to me talk about Sandra for months, patiently, without telling me what to do, which is rarer than it sounds. That evening, she sat at my kitchen table while I showed her the notebook, the folder of camera stills, the certified-mail receipt, the return receipt card signed July 3, and Sandra's note. She read through them quietly, turning pages the way a person reads a file they respect.
Then she put the notebook down and looked at me.
"Stop keeping a diary and start keeping a case file," she said.
She did not say it unkindly. She said it the way Priya says most useful things: plainly, as if the distinction had been obvious all along and I had simply been too close to see it. A diary, she explained, was a record of feelings. A case file was a record of facts, organized so that a stranger could read it and understand, without my help, exactly what had happened and in what order. A diary asked to be believed. A case file did not need to be believed, because it showed its work.
I knew immediately that she was right. My notebook was careful, but it was still mine: my handwriting, my phrasing, my order of emphasis. A stranger reading it would have to trust me. A case file would not ask for trust. It would offer dates, times, signatures, and images, arranged so that the pattern was visible without my narration.
So I reorganized everything. I bought a binder, black, three inches thick, with white adhesive labels, the kind of binder that looks like it means business without trying to. I made tabs. I put the installation papers under the first tab, the maintenance invoices under the second, the correspondence — the certified-mail receipt, the return receipt card, the copy of the letter, Sandra's mailbox note — under the third. And under the fourth tab, which I labeled EVIDENCE, I put the log, rewritten cleanly in date order, and the printed, time-stamped, dated camera stills: May 18, June 7 at 12:40 a.m., and the rest, each one matched to the log entry it supported.
It took me an evening, and then part of another. I checked the cameras as I worked, making sure the system was still recording, still time-stamping, still saving what it saw. I printed stills I had not yet printed, because a still in a camera system is a possibility and a still in a binder is a fact. I cross-checked dates against the log, twice, because an accountant does not file a discrepancy if she can help it.