January 14. Four of the five bundles are files now.
Survey log, two hundred and twelve pages; telegraph log, a hundred and forty-six; stacks log, ninety-seven. With the lighthouse log, five hundred and seventy-three pages. It took three days. What is left is the one notebook, and I decided to do that tomorrow.
Over three days I got used to the work. The hand that lays and the hand that presses move separately, and meanwhile the eye watches the screen. Whether the page that comes up is good or not gets judged half automatically. Tilted, again; shadowed, again. By the third day two hundred pages a day was not a difficult number.
The telegraph log was harder to read than the others. Small writing, dense lines. Messages taken down at night were set down with their hours, and the hours were mostly before dawn. I took it that the man worked nights. Lighthouse keepers work nights too, so there must have been many such posts then.
In the afternoon audit rang. Some of the transfer-roster files handed over last week will not open. It turned out the files are not corrupt; audit's software version is old. The computer room said to convert them.
Converting means opening the original file and saving it again in another format. Saving makes one more file. The original stays and the converted copy is what gets sent, so the same material takes two lines in the list. I converted all twenty-two boxes' worth. It took the whole afternoon, and when it was done the number of files had doubled.
While the conversion ran I looked again at the note I had made. The stroke on the lighthouse log's last page.
I opened the converted copy and found the page. There.
I thought that was a little odd. If there had been grit on the glass during scanning, the grit goes into the file as part of the picture. Change the file to another format and the picture goes with it, so of course it is still there. So there is nothing odd. It took confirming that twice before I let the thought go.
I took the original paper out and looked at the page again. Margin.
What is in the copy and not in the original is also in the copy of the copy. What to call that I did not know. Call it an error and it has to be fixed; call it a feature and it stays. In the log I wrote only note.
With nothing to do while the conversion ran, I sorted the list. By registration order, what had come in over ten days filled two screens. By name, the same material stood side by side, front and back. The original and the converted copy have nearly the same name and differ only in the marker at the end. Set side by side, which was made first is hard to tell from the names alone.
At five I counted the files. Just over two thousand three hundred registered under the archive's name. When I came ten days ago it was in the four hundreds. What has been added since is mostly what was scanned and what was converted.
That the way it grows is different from paper — I thought about that for the first time then.
To make a copy on paper you photograph each page again. A copy of a five-hundred-page bundle is five hundred exposures, and that is three days' work. A file needs one save. Five hundred pages is one. Three days becomes three seconds.
When something gets cheap people do more of it. It was the same at the branch. The year copiers went in on every floor, approval paperwork doubled. Because copies got cheap, not because more paperwork was needed.
So the two thousand three hundred is not more material either. It is the same material in many copies.
Anyone would get that far. What came next caught a little.
When the same material exists in many copies it exists in many places. On audit's screen and my screen and the repository server. Delete one and the others remain. Paper burns and is gone; a file has many places to be burned from.
I was taught that this is a good thing. That is what preservation means. Keeping a thing from being lost.
Only, the stroke on the lighthouse log's last page is now in many places too. Something not in the original sits in several of the two thousand three hundred. Preservation is protecting the original, and what is now in many places is not the original but copies.
I did not think about it for long. I sit here to keep material, not to settle what counts as an original. If it came to that, what audit collected today is a copy too, and the audit is done on the copy. That is how a company runs.
A little before six I rang the computer room. I said a stroke not in the original appears in the scan and in the converted copy. The man thought and asked whether I had rescanned the original. I said I had. Then perhaps the back of the sheet is showing through, he said.
The back is blank, I answered. Then it could be ink bleeding, he said; old paper does that. I said there is no ink at that place on the front either.
He was quiet a moment and then said he would come down next week and look. I said that was fine.
After hanging up I checked the back again. Blank. Held to the light there was nothing.
I wrote the day up in the log. Pages scanned, conversions, total files. Written as numbers the day looked tidy. I wrote it that way for ten years at the branch. If the cash balances the day is over, and if it does not you stay until it does. Here there is no balancing to see. The file count only rises and never falls, so there is nothing to check it against.
Before leaving I wrote down tomorrow's work. Scan the notebook. Around forty pages. Forty pages is ten minutes. After three days of work, a ten-minute job is what is left.
I looked once at the notebook on the desk and went out. A college notebook with no name on the cover and one corner slightly crushed. Tomorrow that becomes a file too. Once it is a file it can be looked up, and once it can be looked up a request can come in.
Who would ever request a thing like this, I thought, and locked the door.
On the stairs it occurred to me to count how far today's converted copies had gone. One on my screen, one on the repository server, one sent to audit. Audit will have opened it and saved it into their own folder. That makes four. When the audit finishes it will be attached to a report, and that makes five.
I made one and counting gives five. Where my part ends was hard to fix. To say the first was mine and the rest grew on their own sounds right, and also does not.