The pictures the world watched from the moon in July of 1969 were dark and murky, and Stan Lebar knew the camera had done better. He had led the team that built it. The television camera on Apollo 11 sent its pictures home in a form no station could broadcast, so on the ground they were converted, and the conversion degraded them. But engineers had also put the camera's own signal, unconverted, onto tape. "It was better," Lebar said. "We knew it was better."
Decades later he and some colleagues went back for those tapes. People would still be playing this footage, he reckoned, in three hundred years. "Those that follow us deserve better than what we had," he said.
It came out in 2006 that nobody knew where the tapes were. Lebar and a television specialist at NASA were among those who spent three years looking, and in 2009 the search reached its conclusion. The tapes had probably been erased, years before, and used again to record data from satellites. NASA had been short of tape.
The lecture before this one sent you home with a file, and had you make one copy of it the same day. This one is about the next fifty years. What do you do with a recording so that fifty years from now somebody can find it, make sense of it, and play it?
Those are three different things, and a recording can fail at any one of them. It can refuse to play. It can play and leave nobody able to say what it is. Or it can sit in perfect order where nobody thinks to look. The search for those tapes ran into all three, and I'll take them in that order: playing, knowing, finding. Each will leave you one thing you have to do and a few you might. What I say about discs, drives and software will date first, so where a year matters I'll give it.
Will it play? The enemy you'd expect is time: tape decays, discs fade. But time isn't what took the moon tapes. Reuse did. The rules NASA worked under let a tape be erased and used again once a program said it no longer needed the data, and that, the searchers concluded, is what was done. "I don't think anyone in the NASA organization did anything wrong," said the specialist who had searched alongside Lebar. "I think it slipped through the cracks." The first lecture had Columbia's oral history office doing much the same in its early years: it typed each interview out, and as a rule recorded the next one over the tape.
Then there's the machine, because a recording can outlast whatever plays it. In 1986 the BBC made a new Doomsday Book, a portrait of Britain on two video discs that needed a special computer. The first Doomsday Book, an inventory of England, was nine hundred years older.
By 2002, a newspaper reported, only one of the two Doomsday Books could still be read, and you know which. The computers were obsolete. The old book was in the Public Record Office, open to anyone who could read and had the right credentials. One of the people trying to rescue the discs said, "We're lucky Shakespeare didn't write on an old PC."
And there's decay itself, where you'd like a number of years and what you get is a spread. A guide to tape written for libraries and archives in the 1990s gives it ten to thirty years, and compares figures like that to a car's miles per gallon: your mileage may vary. For the drive inside a computer I have no lifetime in years. Doug Boyd, the Kentucky oral historian from the lecture before this one, says to assume that it will fail.
What to buy keeps changing as well. In 2012 Boyd was moving people off discs and onto hard drives, and a year later a writer at the Library of Congress said what still holds: no one kind of storage is best. So what protects a recording is a habit, and the habit is copies.
You made the first copy the day you recorded her. The first lecture called the recording the original, as against the typed page. Here the word narrows to the file as the machine made it: nothing gets trimmed from it or thrown away to make it smaller, and any work is done on a copy. Now play the copy. Boyd checks his copies with a tool that compares the files exactly. If you have none, I'd do this, and it's what playing a copy means from here on: see that it runs to the same length as the original, and listen to its beginning, its middle and its last minute.
Make a second copy on a separate drive, play it, and take it out of the house, to somewhere that is still yours: a drawer at work, say.
Then she hears the recording, before anyone else does, and I'd make it that week. That was a promise from the sixth lecture, and it's when she says what's kept, what's closed until a date she names, and what's cut for good. If she'll have a part closed and not cut, nothing in your files has to change, and the part will still be there if she, or her children, think again.
A cut is the one time the original changes, and no guide walks you through one, so this is how I'd do it. Make it on a copy. Save that copy in a format that throws nothing away, because the Oral History Association warns that a compressed file, saved again, is compressed again. Check it by ear, and let her hear it. Only then does the cut version replace every other copy, the recorder's included. And cut for good has to mean it: the passage also goes from wherever your machines keep what's been deleted, and from anything that copied the file by itself.
The third copy can now go to live with the person who will hold the recording after you. That makes the one thing you have to do: three copies, in at least two places, each one played. The guides differ on the number, from at least two to at least four, and three is my choice among them, following Boyd. They agree that the copies shouldn't all be in one building, because of fire and flood, and drives that share a house share a place. Don't hurry to clear the recorder. Baylor's manual would have the files safe in at least four other places first, and at the least I'd wait until the three have been played. Clearing it sooner is how a family reuses its tape.
This is where somebody says that the phone backs itself up. One state archive grants that a backup may be a good copy, depending on how it's set up. It then warns against trusting automation you don't fully understand, and Boyd's warning is that an automatic backup copies a damaged file as faithfully as a good one. So I wouldn't count it among the three, which are copies you made and played yourself. An online service you put the file on yourself is one of the things you might do, though the Oral History Association suggests it only for material that isn't sensitive, so not before she's heard it. It is never the only copy. Boyd says so, and a retired forklift driver in the north of England could say why.
He had kept his late wife's voice, the greeting on their voicemail, for more than a decade. Then his phone company upgraded the line, having assured him the recording would be saved, and wiped it. The company put ten engineers on the job, and they got it back. He had tried, earlier, to make a copy of his own. "All I ended up with were blank tapes," he said. A copy you haven't played isn't yet a copy.
After that it's a calendar. The Library of Congress's advice to people keeping their own recordings is to check each copy at least once a year, and to copy them onto new drives about every five. I'd keep the old drives until the new copies have played. If the format you recorded in is falling out of use, add a copy in the new one, play it, and keep the original.
Do all that and the recording should play. It may still be a mystery to whoever plays it. Late in NASA's search, old documents showed that a laboratory near Baltimore had recorded the camera's signal as well, on a couple of tapes of a different kind. The searchers hunted up a machine old enough to play them. They expected to have to bake those tapes before they would run, so they bought a vegetable dehydrator. I'm reporting that, not recommending it. The laboratory's archivist did turn up tapes of the right kind and the right years. They had no labels, and when they were played, they were blank. Were those the moonwalk? There's no way to know.
So a recording has to say what it is. Yours does already, on the inside, because you said the slate in its first minute: the date, the place and the names. The outside is the name of the file. Whatever the machine called it, change it to her full name and the date. Give every copy that name. I'd add a word that says which it is: the original, or a copy made for the family.
Then the index. An index is a list of what she talks about and where in the recording it comes: the time, a few words about the topic, and the names spelled right. In Santa Fe, one January, a husband and wife sat down at the kitchenette table in a StoryCorps trailer to record each other, above all about the vacation on which they had both come down with bubonic plague. They talked for forty minutes. Off to one side sat a StoryCorps staff member, writing notes and keywords against the time on the recording. By 2013 StoryCorps had recorded almost a hundred thousand people, more than its staff could play back, let alone type out, and those notes were the way in.
You were busy listening, so yours gets made afterwards. Soon afterwards, says the manual from Baylor University. I'd say within the week, while she can still be asked. At the top go her name, yours, the date and the place. Then play the recording, and each time the subject changes, or every five minutes if it doesn't, write down the time and what she's talking about. A line might read: twelve minutes, the lodger in the front room, his name spelled as she spelled it for you.
The other thing you might write is a transcript, which is the recording typed out, word for word. Baylor calls an index a satisfactory alternative to one. A writer at the Library of Congress, describing an indexing tool Boyd helped create, goes further: an index finds what a search of a transcript can't, because a person's judgment made it. Somebody can talk for two hours about life under segregation, the writer points out, and never say the word. So the index is the one thing you have to write about what she said.
A transcript has its uses. It can be skimmed, she can read it and correct it, and paper keeps. Against that, Baylor reckons about ten to twelve hours to transcribe an hour of talk. The historian Linda Shopes says that typing an interview out takes the oral out of oral history. It loses the voice, which is what Jacqueline Kennedy's daughter, in the lecture before this one, got from the tapes and not from the page.
A page can also hold what was never said. Boyd tells of a transcript of somebody's account of the Second World War that was checked against its tape. In the transcript, he says, were the words "an amazing Italian who was fast and quick." The tape had an amazing battalion. Once a thing like that is typed, it tends to be taken as fact.
Software will now do the typing, and it makes mistakes of its own. In 2012 Shopes wrote that no workable system existed for the way oral history narrators talk. In 2023 researchers tried a leading program on recorded interviews, a sentence at a time. In roughly one sentence in a hundred it added a phrase, or a whole sentence, that nobody had said, and more often where the speaker left long pauses. Somebody retelling Cinderella got as far as the dance with the prince, and the program added, "Thank you for watching!" Some of those speakers had aphasia, which leaves long gaps in speech, and the authors think the same may go for the very old. You've been counting to ten in her silences.
A study reported in 2026 ran a later version of that program over an archive of Holocaust survivors' testimonies at Yale, most of them recorded in old age. It got about eighty-five words in a hundred right, it too sometimes wrote what nobody said, and the authors would have a person go over the names and the dates. So what software gives you is a draft. Read it against the recording from beginning to end, and then let her read it and add her corrections.
Where the software runs matters. I'd find out whether your recording leaves your computer to be typed, and whether whoever types it keeps a copy. If they do, the question is whether she agreed to that. I'd print the index, since paper needs no machine to read it, though the paper Shopes calls the most stable medium of all is the archival kind.
The third failure is the one NASA met first, and it cost them those years of searching. The searchers' count was about forty-five tapes, sent into storage among several hundred thousand. They kept going back to one government warehouse, which Lebar compared to the one at the end of Raiders of the Lost Ark.
Nothing has to be wrong with a recording for it to be lost, and the remedy is a sheet of paper. The index was about what she said. This is about where the recording is. The Library of Congress's advice is to write a short description of what you have, and to put a copy of it with your important papers. Yours says what recordings there are, what the files are called, and where each copy is. I'd add more. It says how the next person gets into anything that needs a password or an account. If a cut was made, it says so, and when, and that she asked for it. It's dated, and written again whenever a copy moves. Put a copy of it, with the index, on every drive beside the recording. If she signed a release, keep that with it, since it records what's closed and until when, and whom she wants to hold the recording after you. Tell that person where all this is.
The family comes into it here. Densho, an archive that writes for families, says to share with relatives and ask them to add notes of their own. I'd share the recording as well as any page, since every relative's house is another place, and I'd hand over the file itself, on a drive. Give them what she agreed to, with any closed part left out of a copy made for the purpose, and put nothing online unless her yes covered that. Where a part is closed, the whole original stays with you and with the person who holds it after you, and nobody else. Otherwise let the original go round as well, and say which file it is.
Saying which file is the original matters, because of a comparison that is mine to answer for. At Columbia what was kept was the typed page, and the tape, as a rule, was used again. What's left of the first moonwalk is the broadcast, the murky conversion. Twice, then, the version people could use outlasted the original. I'd expect a family to do the same with a clip on a phone, unless everybody knows which file is the original.
If the recording ought to outlast the family's interest in it, there are archives: a historical society, or a library. Don't mail it to one. The advice I have is American, from the Society of American Archivists. Ask first, because an archive can't take everything and may say no. If it says yes, it will want the recording identified and every copy accounted for. It will most likely ask you to sign a deed of gift, which makes what you give the archive's property. Copyright is separate, and archives often ask for that as well. So settle it with her before you sign, and have her sign too: by Baylor's manual, her signed release goes to the archive with the recording. Anything closed is agreed with the archive first: the Society says archives will agree to reasonable restrictions for a limited time. I'd call that a promise as far as the law allows.
It's fifty years on, and somebody wants to hear her. Can they find it? There's a sheet of paper with the family's papers that says where the copies are, and somebody was told. Do they know what they've found? Her name and the date are on the file and are spoken in its first minute, and an index says what's in it and where. Will it play? It has been opened every year and moved every five, it was never the only one, and it's the original and not somebody's clip.
Two things I can't give you. One is a lifetime in years for anything on your shelf. The other is a ruling on whether to type the whole thing out, where the people who do this for a living still differ.
The tape this course opened on belongs here. Dave Isay recorded his grandmother and her sisters when he was a boy. At seventeen he went to find the tape and couldn't, and at forty-two he told a reporter he still searched his parents' house for it. He doesn't say what became of it, only that he can't find it. That's the third failure, the one with no machine in it.
What he built afterwards was StoryCorps, and it has this lecture inside it. That couple in Santa Fe went home with a copy. Their recording went on to the Library of Congress, which keeps its copies in more than one place, and the notes and the signed forms went with it.
What a family can manage is three copies, an index and a sheet of paper. They're for the seventeen-year-old who goes looking.