February 1. A circular came down: the archive's work is being folded into the computer room.
The content was simple. The physical material has all gone to the outside repository, so no one needs to be stationed in the basement archive. Requests come in electronically and are handled electronically, so the keeper can do the same work from the computer room. The effective date was February 8.
After a month I packed again. There was not much to pack. One notebook, a pen, the calculator I brought from the branch. I thought about throwing out the scrap in the drawer and left it. Whether anyone will use this desk after me I do not know, but if someone does, they will open the drawer too.
The problem was the box. It had sat under the desk since the call to personnel a month ago, and in that time everything inside it had become files.
The box is on no list, so I could not write a withdrawal form. That the disposal form has no line for a wooden box I confirmed last month. Left behind, it becomes a thing not on any list sitting in an empty room, and someone cleaning will throw it out eventually.
It is a thing that could be thrown out. What was in it is all files now, and the files are in many places. The paper has become the copy side.
Even so I decided to carry it up. If I had to give a reason it would be that it is the original of old material, and that is not the exact reason either. Exactly put, throwing it out did not sit well. It is a thing that touched my knee for a month.
In the week before the move I handled twenty-three requests. Nineteen closed as files and only four went to the repository for physical material. A physical request takes two days round trip and a file takes thirty minutes. Which one people will ask for does not need asking.
Where this work is heading was roughly visible too. As physical requests fall, someone will say the repository costs too much, and then someone will say to reduce the physical copies of whatever has been scanned. Not mine to decide, but the direction is clear enough.
On February 8 I emptied the archive.
I looked round the room one last time. Six rows of bare shelving, three desks, three terminals. The terminals stay as assets, they said. I dropped the power and left them. As it dropped the screen brightened once and went dark. I took that for the normal thing.
I turned off the lights and locked the door. No one is stationed on the second basement floor now. I returned the key to facilities, and if a physical request comes in facilities will open it. A room used for a month and some, and handing the key over took three minutes.
The computer room is on the sixth floor. It has windows, it is bright, and twelve people work there. The server room is next door with its door always shut and the sound of machines running inside. The sound holds at the same pitch all day. It is not like listening to pipes cooling in the basement. That was a cooling sound and this is a running sound.
My seat is right beside the server-room door. The box went under the desk.
I rested once on the stairs carrying it up. From the second basement to the sixth floor is eight flights. It is the box I tried to move a month ago and gave up on because it was heavy. Then I pushed it with a knee; this time I carried it up eight flights. The paper inside is the same paper, and it counts as the copy now — so I carried a box of copies up eight flights.
The same work in a different seat feels like different work. Taking a request and handling it is the same. Only, in the basement requests were two or three a day and here they run past ten. In the computer room people say things in passing instead of ringing. It is faster.
The people in the computer room do not see material as material. They see capacity. This month's growth, what percentage the disk is at, when it will need expanding at this rate. At first the way they talk was strange and after a few days I was talking that way too. I now know how many megabytes a hundred and eighteen pages of lighthouse log come to.
About a week in, I had reason to look at the night job log. A file of audit's was not coming up in the index and they asked me to check.
The log records what the server does at night with the hours. It was a screen I had not seen. The person beside me showed me how to open it.
There were several jobs. Backup, cleanup, indexing. Each with a start and an end.
The indexing job started at three.
I looked at the line for a moment. Three in the morning. Running heavy jobs when the server is idle is the obvious thing. Backup at two, cleanup at half past, indexing at three. Someone set it that way.
The hours from the telegraph log came to me after that.
The messages in that log were mostly between three and four in the morning. I read that as the man working nights and I still read it that way. That the hours a man worked at night sixty years ago and the hours a server works at night now should overlap is because those hours are the quietest. People and machines both do heavy work when it is quiet.
Put that way there was nothing odd. I found audit's file, confirmed why the index had skipped it and replied. A special character in the file name had made the indexing job skip it. Renamed and re-registered, it came up the next morning.
Before closing the log I looked at one more thing.
The indexing log records how many files were indexed that night. Ours, other departments', the archive's. The archive line was still there. There is no one in the archive now and the line remains. If a department code is live, whatever is registered under it goes on being indexed.
That night eleven files were indexed under the archive code. The night before it was nine, and the night before that twelve. Set as three lines, something was being registered every day in about that number.
I registered nothing in the archive. I emptied that room last week and dropped the terminals' power.
I asked the person beside me. He looked at the screen a moment and said that what comes up from the repository registers under the archive code. The physical material is there, so it was set that way.
That settles it. I closed the log.
After closing it I thought briefly about checking what had come up from the repository that day. Seeing eleven items as a list would have done it. Two screens.
I did not. The reason was nothing much. It was nearly six and three requests were still waiting. If it needs checking it can be checked tomorrow.
I did not write it down to check tomorrow, either.