Kipple

Transcript

1. Julian Pankratz

   This is Audio Logs. I’m Julian Pankratz, and today I’ve come north to Vancouver, to the Lazarus reading room, where anyone can walk in off the street and run software that hasn’t had a working home in thirty years. Lazarus has spent thirty-two years bringing back the software, records and digital lives that everyone else had written off. I’m with two members of the team: Miranda Lawson, who leads the program, and the archivist J. R. Isidore, whose essay on digital clutter has become something of a cult favorite around the office. Welcome to you both.

2. Miranda Lawson

   Thanks, Julian.

3. J. R. Isidore

   Hello. Thank you. I don’t usually do this sort of thing.

4. Julian Pankratz

   You’ll be fine. Just think of the microphone as a very old disk drive.

5. J. R. Isidore

   Then I’d be much more careful with it than you are. Sorry. That came out wrong. You’re holding it fine.

6. Miranda Lawson

   He’s right, though. You’re holding it like a wine bottle.

7. Julian Pankratz

   Thirty seconds in and I’ve already been appraised. Your essay is called “Kipple.” What is kipple?

8. J. R. Isidore

   It’s a word an old neighbor of mine used. It means the useless things that pile up when nobody is paying attention. Junk mail. Empty matchbooks. Old cables. Nobody brings them in on purpose; they just arrive. And his theory was that kipple drives out everything else. Leave a room alone long enough and you can’t find anything that matters in it.

9. Julian Pankratz

   He sounds like quite a character.

10. J. R. Isidore

    He was. He lived in a big building that was nearly empty, and he could tell you which rooms had gone under, and in what order. I live in one like that now, actually. Lots of empty floors. I keep my own apartment clear. That’s, um. That’s the rule.

11. Julian Pankratz

    And data is the same?

12. J. R. Isidore

    Data is worse. Physical clutter at least takes up space you can see. When we review an organization’s storage, usually two-thirds of the files have never been opened since the day they were made. The important records are in there somewhere, in four slightly different copies.

13. Miranda Lawson

    Three point four, on average. J. R. rounds up.

14. J. R. Isidore

    It feels like four when you’re the one looking for the right one.

15. Julian Pankratz

    J. Bone was on the show a few weeks ago, and he described you, Miranda, as scary precise.

16. Miranda Lawson

    He meant it as a compliment, and I took it as one. And J. R.’s point stands. People assume that keeping everything is the safe option. It isn’t. Every useless file is something that has to be migrated, secured and searched through. Clutter is how archives lose things without ever deleting anything.

17. Julian Pankratz

    I should confess something here. I have every draft of every poem I’ve ever written. Spread across three laptops. Two of which no longer turn on.

18. J. R. Isidore

    Do you know which one has the final draft?

19. Julian Pankratz

    There’s a final draft?

20. Miranda Lawson

    That’s the whole problem in one sentence. Nobody decides to hoard. It just accumulates, and one day the copy you need is on the laptop that won’t turn on.

21. Julian Pankratz

    Miranda, earlier this year you published a report called Seepage, about storage that’s been asked to hold too much. Is that kipple’s evil twin?

22. Miranda Lawson

    Its consequence, more like. We looked at three thousand devices we’d recovered over eleven years. Drives kept more than ninety percent full for over a year were six times as likely to show corruption as drives kept under seventy. And it’s mostly quiet corruption. A drive that dies, you replace. A drive that silently changes one record in a thousand, people go on trusting for years. In about one overloaded device in ten, pieces of one record had leaked into the record next to it.

23. J. R. Isidore

    Kipple fills the room. Seepage is what happens when the room’s too full and things start getting into each other.

24. Julian Pankratz

    That’s going on a mug. Now, eagle-eared listeners may have noticed that the archive’s search had a wobble last month.

25. Miranda Lawson

    Three hours and ten minutes, on the nineteenth of August. We moved the search index to new storage, and while it moved, searches came back incomplete. Nothing in the archive itself was touched. Every object we hold has three copies: two here in Vancouver, on different kinds of storage, and one at a partner site a long way away. And there’s a service we call Vigil that re-reads and re-checks every one of them, continuously. A full pass over three point one petabytes takes about eleven weeks, and then it starts again. After the move, Vigil made an extra pass over everything the move had touched. No mismatches.

26. Julian Pankratz

    Has Vigil ever actually caught anything?

27. Miranda Lawson

    All the time. A copy stops matching its checksum, it’s taken out of use, the other two are checked, and a good one replaces it. For us to really lose something, the same damage would have to turn up in all three copies before Vigil next looked. In thirty-two years, that hasn’t happened.

28. J. R. Isidore

    Our new CTO came up to see Vigil the week he started. He stood in front of the storage for a while, very quietly, and then he thanked it. I liked him immediately.

29. Miranda Lawson

    We all did. Though I’d like it on the record that Vigil was working perfectly well before anyone thanked it.

30. Julian Pankratz

    And the move kicked off a much bigger job.

31. Miranda Lawson

    Re-cataloguing. Every record is being brought up to the fourth edition of our descriptive standard: description reviewed, checksum re-verified, rights checked, by a person, before it goes back into the online catalog. It will take a while. Everything stays available in the reading room in the meantime.

32. J. R. Isidore

    It’s my favorite kind of work, honestly. It’s like walking through every room in the building again, with the lights on.

33. Julian Pankratz

    Miranda, Lazarus is known for bringing things back. What’s the hardest recovery you’ve done?

34. Miranda Lawson

    Technically? A set of drives that had been through a fire. But the hardest ones are never technical. They’re the cases where a family wants something back, a person, really, in the form of their messages and photos, and you have to be honest about what can and can’t be restored. I’ve spent my career bringing back things everyone said were gone. You learn to be careful about what you promise.

35. Julian Pankratz

    I want to come back to the families. But tell me about the fire first.

36. Miranda Lawson

    Kingsmere Public Library. A small library that burned down in 2016. Their four server drives were pulled out of what was left of the server room, blackened and partly melted. The first rule with a drive like that is that you don’t power it on. Ever. You’ll destroy whatever’s left.

37. Julian Pankratz

    So what do you do instead?

38. Miranda Lawson

    You open each drive on a clean bench and clean the soot off every platter by hand. Then you move the platters into matching donor drives, because the original mechanics are finished. And then you discover that not one of the four can be read all the way through.

39. Julian Pankratz

    That sounds like the end of the story.

40. Miranda Lawson

    It very nearly was. But the library’s server had spread its data across all four drives, with some redundancy, so that it could survive losing one of them. We hadn’t lost one. We’d lost different pieces of all four. So it became a puzzle: working out which gaps on one drive were covered by what had survived on another. It took months. In the end, between them, the four drives held enough to rebuild the whole database.

41. J. R. Isidore

    Our catalog calls that fidelity level reconstructed. The result is whole, but no single original ever held it. It’s my favorite category.

42. Miranda Lawson

    It is not a category anyone aims for, J. R.

43. J. R. Isidore

    No. But it’s a nice thing to be able to say about something.

44. Julian Pankratz

    And what was in it, in the end?

45. Miranda Lawson

    The catalog of about sixty thousand books, a lot of it local history held nowhere else, and eighteen years of lending records, which the library anonymized before they gave them to us. The books themselves were gone. That catalog was the only complete list of what the town had lost, so at least they knew what to look for. The whole story is on the archive site.

46. Julian Pankratz

    Now, the families.

47. Miranda Lawson

    We’re often asked to help with a phone or a laptop that belonged to someone who has died. We only ever work with the consent and the credentials of the owner or their estate; we don’t break into anything. And very often we can bring back the photographs, the messages, the voice memos. But what people sometimes want, underneath all that, is the person. More and more services now offer to build a replica that talks like them. We won’t. Nothing in this archive is ever used to make anyone talk again.

48. Julian Pankratz

    That’s at the heart of the Half-Life brief you published last month.

49. Miranda Lawson

    With Herbert Schoenheit von Vogelsang, who has run a memorial archive in Zürich for twenty years. His phrase is visiting, not resurrecting. You can read the letters, listen to the recordings, remember. Nothing new is generated in their voice. And the brief has a quieter finding, too. We traced two thousand estates. Roughly half of a person’s accounts become inaccessible to their family within five years of their death. Almost none of that loss is deliberate. It’s simply nobody’s job to stop it.

50. Julian Pankratz

    And for a listener who’d like to make it somebody’s job?

51. Miranda Lawson

    Write one page. Who should have access to your accounts, where your files are, and whether you’d like to be remembered, forgotten or simply left alone. Keep it with your will, and tell someone where it is.

52. Julian Pankratz

    Before Frothly, you ran a restoration project for an organization you prefer not to name. I’m professionally obliged to ask.

53. Miranda Lawson

    And I’m professionally obliged to keep preferring. It was very ambitious and very well funded, and it taught me the difference between restoring something and making it into what someone else wants it to be. Lazarus restores. That’s all.

54. Julian Pankratz

    J. R., how did you come to archiving?

55. J. R. Isidore

    I used to repair animals. Electric ones, mostly, for a pet hospital. You learn to tell what’s worth fixing. Once I got one wrong. I thought it was a machine and it wasn’t. After that I wanted to work somewhere the point was to keep things, not to judge them too quickly.

56. Julian Pankratz

    That’s quite a story.

57. J. R. Isidore

    I think about it a lot. It’s probably why I’m careful about deleting anything. But careful isn’t the same as keeping everything. You ask four questions. Is it unique? Is it evidence of something? Would anyone want it in fifty years? Can we keep it readable that long? If the answer to all four is no, you let it go, and you write down that you did.

58. Miranda Lawson

    That last part matters. Deliberate deletion, with a record, is part of preservation. It’s the opposite of losing things.

59. Julian Pankratz

    Give me an example of something that passed.

60. J. R. Isidore

    When my old hospital closed its repair department, they gave us their service firmware. Sixty-one models. Cats, dogs, horses, a goat, an owl. And the diagnostic programs the technicians used to tell a failing animal from one that was just unhappy. Nobody else had kept those. So, unique, and evidence of a whole kind of work that’s mostly gone now. I catalogued it myself. There was a toad, too. The technicians kept it on a shelf, as a curiosity. We read that one as well.

61. Julian Pankratz

    And something that failed all four?

62. J. R. Isidore

    A print server, from a company that asked us for help. Forty thousand log files, and every one of them said, more or less, paper low. Not unique. Not evidence of anything except a printer. Nobody in fifty years. And every one of them would have needed migrating, forever. So we let them go, and we wrote down what they were and who decided. Me, that time.

63. Miranda Lawson

    So if a historian of office printing ever turns up angry, they’ll know whose desk to come to.

64. J. R. Isidore

    I’d apologize. I’d mean it, too.

65. Julian Pankratz

    There’s another side to this, though. Some of the most moving things you hold are things somebody kept for no sensible reason at all.

66. Miranda Lawson

    Our very first record. It’s before my time, but everyone who works here hears it in their first week. March 1994, a borrowed church hall here in Vancouver, the first Lazarus rescue day. Someone brings in a floppy disk that’s spent nine years in a kitchen drawer. It reads on the third attempt, on a drive one of the volunteers had repaired that morning. On it there’s one document: a letter his grandfather wrote to him in the winter of 1986 and never sent. They printed it that afternoon and handed it to him.

67. Julian Pankratz

    Oh, that’s lovely.

68. J. R. Isidore

    We keep the letter, but we don’t show it. It’s his. The word processor it was written in, anyone can use.

69. Julian Pankratz

    And this summer you reached the forty-one thousandth title.

70. J. R. Isidore

    Space Lunch. A game an eleven-year-old girl wrote in 1989, on a cassette her parents kept for thirty-seven years. You steer a lunchbox through an asteroid field to an astronaut who’s forgotten his sandwiches. Three levels, a high-score table and a title tune. She’s an engineer now, and she sent it to us herself.

71. Miranda Lawson

    The tape had stretched, so we sampled it as audio at reduced speed and repaired the damaged blocks by hand, from the waveform. She’s released it for anyone to play.

72. J. R. Isidore

    I’ve played it quite a lot. I’m not very good. The asteroids on the third level are very unkind.

73. Julian Pankratz

    So the first record and the forty-one thousandth are both things a family put in a drawer and couldn’t bring themselves to throw away.

74. J. R. Isidore

    Yes. And that’s what I’d most want people to hear, I think. Nobody keeps a cassette for thirty-seven years by accident. That isn’t kipple. That’s the opposite of kipple. Kipple is what arrives by itself. Keeping is something you do.

75. Julian Pankratz

    Any advice for listeners staring at a hard drive full of kipple?

76. J. R. Isidore

    Don’t try to win. You won’t. Just keep one room clear. Pick the folder that matters most, sort it, and keep it sorted. That’s enough. That’s more than most people manage.

77. Julian Pankratz

    While I have you both, I have a shoebox of cassettes from my touring years that I’ve never dared to play.

78. Miranda Lawson

    Good. Don’t. If something old and precious might be failing, the best thing you can do for it is nothing, carefully, until someone with the right equipment can read it once. A lot of the damage we see was done the week before it reached us, by someone trying to help.

79. J. R. Isidore

    Are they unique?

80. Julian Pankratz

    Tragically, yes.

81. J. R. Isidore

    Then they pass the first question. You should bring them in.

82. Miranda Lawson

    And whatever you do keep, keep three copies, on two kinds of storage, with one somewhere other than your house. Two copies in the same house are one fire away from none. I say that as someone who has cleaned soot off platters by hand.

83. Julian Pankratz

    Miranda Lawson, J. R. Isidore, thank you both.

84. Miranda Lawson

    Thank you, Julian. Bring the shoebox.

85. J. R. Isidore

    Thank you. That was nicer than I thought it would be. I think I’d do it again.

86. Julian Pankratz

    That was Audio Logs. J. R.’s essay, Kipple: The Slow Entropy of Unused Data, and the Half-Life brief are in our research library, and the story of the Kingsmere drives, along with some forty-one thousand others, is on the archive site. We’ll be back in two weeks.