The end when it came, came slowly... cracks slowly appeared in the world's economy as the endless quarterly search for increases in growth started running against the hard limits of the world's actual physical resources.

It wasn't obvious to many for a long time; the governments said unemployment was down, because they stopped counting people who couldn't get a job; GDP was up, because we counted increasingly abstract financial products selling back and forth as productive; gas prices went down, because fewer people could afford a car. Some countries had rebellions, uprisings, and civil wars, but they were always far away. Political discourse had deteriorated to ad hominem attacks and compromise was no where to be found, but hadn't politicians always been that way. You couldn't buy high quality products anymore no matter how much you were willing to pay. Shortages of goods were more common; but there was always a plausible reason; a strike, a flood, an accident at the production plant. Families were working twice as hard, yet standards of living were still going down.

As the cracks got wider some people started to notice.

As they saw the signs and recognized the decline of a civilization they decided to make plans to retain and use societies accumulated knowledge, thus came the question:

How best to store a very large amount of information given the following constraints:

  • Capable of storing a large amount of information, say about 1 LOC (The Library of Congress print collection is equal to ~15 TB by some estimates)
  • Technological production and supply chains are starting to break down, replacement parts more complex than can be manufactured by a hobbyist or group with simple tools are increasingly unavailable and expensive. Soon (~50 years) no complex chips, hard drives, or other high precision manufacturing replacement parts are going to be available (the possibility of stockpiling backup parts is okay but should work within the other constraints and should think about shelf life of backups)
  • Electrical grids and grid scale power sources are failing, some locations have no electricity others have limited power with frequent blackouts. Some renewable electrical sources are available but provide much lower capacity than currently is normally available (Solar PV cells are complex to manufacture and will be unreplaceable after they wear out in ~40-50years)
  • Must be a durable medium capable of lasting a long time under existing conditions (500+ years). Copying information or manufacturing replacement parts is okay to meet this requirement, but should fit in the other limitations.
  • Must be human readable with maintainable support equipment; full read/write capabilities a plus.
  • Storage method should allow indexing and cross references, the archive is useless if you can't find anything. Trained specialists are okay to meet this requirement (I always liked librarians).
  • Redundancy, capable of multiple copies being made and distributed to allow for loss of data at one location. Speed and accuracy of copying would also be helpful (it would take a lot of people to hand write and copy the library of congress and would require some gifted artists/draftsmen to copy some non-text resources).

Some ideas I have had already considered to be good possibilities:

  • Books, including mechanical printing press technology (The Silo series used a specially designed encyclopedia)
  • Microfilm and readers requiring only optical magnification to read (copies and new info can be supplied with some pretty simple photographic chemistry
  • Punch cards or other simple electro-mechanical computer storage methods
  • A distributed system (regional library or a Saint Leibowitz religious order)

Which data storage method would best meet the requirements?

  • $\begingroup$ Why wouldn't they be able to make microchips? Will they run of of sand? It's easier to make microchips than complex mechanical machinery or optical devices. If by electro-mechanical computer storage you mean core memory that was incredibly hard to make: microchips are child's play compared to magnetic core memory. $\endgroup$
    – AlexP
    Dec 22, 2016 at 17:50
  • 1
    $\begingroup$ @AlexP chips are a lot more than sand, you need very chemically pure materials and complex precision machinery, which all relies on a complex international supply chain to provide. A global economic collapse would destroy the supply chain. You wouldn't have any newly manufactured computer memory of any type. By SIMPLE electro mechanical I was thinking of more a read write mechanism utilizing something physical like a punch card rather than the being the memory itself. $\endgroup$
    – Josh King
    Dec 22, 2016 at 18:18
  • 1
    $\begingroup$ How pure the chemicals need to be depends on the kind of chips you want to make. It's one thing to make a high-density SSD and quite another to make a 2.5 MHz Z80. In the dark days of Communism the Eastern Block countries made decent computers without much if any access to any kind of global supply chain. Do not underestimate the capacity of engineers to adapt. Losing iPhones because of the disruption of global supply chains is possible, but completely losing electronics is unlikely in the absence of total civilzation meltdown. $\endgroup$
    – AlexP
    Dec 22, 2016 at 18:30
  • $\begingroup$ Heavily laminated cardstock? $\endgroup$ Dec 22, 2016 at 18:48
  • 1
    $\begingroup$ Use really, really, really small font. $\endgroup$ Dec 22, 2016 at 19:27

3 Answers 3


The real issue in the beginning is going to be uncertainty. No one really knows how far civilization is going to slide. Are we headed back to the Bronze Age or is this just a temporary disruption of the global infrastructure?

If I am convinced that civilization has a reasonable chance of ending, as we know it, to the point I am planning how to preserve knowledge I am going to recommend a paper library. We already have reliable indexing systems. We will be able to make quality paper, ink and printing machines a long time after Apple can't ship another iphone. You could store information in something denser. Pre-industrial microfilm is an interesting concept. Electro-mechanical punch card reader with arc lamp display would be super cool. But with the world going to heck and nobody knows how bad it's going to get, I would start printing like crazy.

How much paper are we talking here?

15TB = 15 Trillion printable characters

250 words per page (https://www.quora.com/What-is-the-standard-amount-of-characters-for-one-page-typed)

average word length = 5

Pages = 15 trillion / 5 / 250 = 12 Billion pages

I found a ream of 500 sheets of acid free archival paper online for $20.

Cost = 12 * 10^9 / 500 * 20 = 480 Million dollars

That's just paper. You'll need ink and binding materials and copy machines. You'll probably need to pay for the necessities of your librarians as they make the copies and indexes. Say $1-2 Billion per copy, and you'd want at least two as far apart as practical.

The biggest issue is going to be organizing the people needed. I mean, you've got to find people who aren't scrambling just to survive, who have the means to commit time to this endeavor.

Maybe get these guys involved: http://longnow.org

PS. Your opening description gave me chills. I'll be hiding in the basement if anyone needs me.

  • 1
    $\begingroup$ Note: That 250 words/page wastes a lot of paper. 700 words/page is more like it if your prime objective is minimizing the cost--and that's before you consider reducing the font size. Also, double-side it. On the flip side, I would be inclined to use tyvek paper--waterproof and it won't tear. Counting all these changes, add about 50% to your cost. $\endgroup$ Dec 23, 2016 at 3:26
  • $\begingroup$ Of course most of your material is already in books, and an archive already exists, plus of course you get a pretty big discount when you buy paper my the billions of sheets. $\endgroup$
    – John
    Dec 7, 2021 at 1:46


I would like to offer a digital solution. There is a DVD-like format called MDISC which has a theoretical lifetime of 1000+ years. The rest of my answer will be in support of this format.

The problem then comes to the reading/writing of the media. Here's where I haven't done enough research, but I know what the requirements would be, and I think it would be possible to construct a device that can reliably operate over millennia.

  1. A power source
  2. Computing power (a processor, RAM, etc...) that's durable
  3. An MDISC reader/writer
  4. Human interface device(s)

In your world described above, there can be forms of renewable energy, and I would imagine one computer could run on such a power source, so #1 is covered. This could be implemented with Solar, Wind, even a Water wheel. A water tower can be an energy storage solution as well (instead of deteriorating batteries).

Numbers 2-4 are where I'm not sure as much. Here's what I do know:

Military computers are designed to last. They take into account impacts, water damage, rust, particulates, and even radiation. Satellite computers use radiation hardened circuitry that resists radiation interference. They are also designed to tolerate extreme temperature environments. Hence I believe #2 can be overcome with some design work. This of course assumes it's not too late to acquire such technology.

Numbers 3 and 4 are more difficult:

Current MDISC technology seems to be used with our modern DVD/CD reader/writers. I haven't looked very hard, but it doesn't seem like a durable (withstanding hundreds of years) reader/writer currently exists. I suppose the assumption of MDISC is that we'll still have those around in the millennia. That's the problem of #3.

I would guess that similar Military designs could be applied to a keyboard and a screen that would make them durable enough for a long amount of time. Ditch the mouse idea altogether. Perhaps create a custom OS that's optimized for keyboard-only use. This would help the case for #4.

Of course then this durable computing machine would have to be replicated, along with the ~700 MDISCs (15,000GB / 23GB = ~653 discs and let's be sure to have some replacements) and then distributed to the several "knowledge safe-houses" that would be set up. This again assumes it's not too late to acquire the necessary components.


  • Much smaller (and portable) than an IRL paper-and-ink Library of Congress. A simple indexing system could be devised to keep the discs in order in their special fire/shock-proof case they would be kept in. I would imagine one person could carry the system in a backpack and/or rolling suitcase.
  • Did I mention fire? This solution would be durable against fire, a common post-apocalyptic (no matter how slow) disaster.
  • Easy interface. The most complicated part would be the organization of the discs. We'll assume the user can read and follow onscreen instructions.


  • After re-reading the OP, I found I made an assumption that technology would still be available at the time of the creation of this system. If that is not the case, then this system would indeed be difficult to implement, if not impossible.
  • I have my doubts about how well a screen would last. Perhaps the design could allow for interchangeable parts? Yeah it should.

Another note:

We might assume that some of the higher-ups still have access to some types of technology. They would likely be the only people even worried about implementing such a system. Not to mention people who have already implemented systems today.

I don't know of an example, but I bet someone in the world has compiled human history and knowledge and already thought up and implemented such a system. In a "Slow Apocalypse" as has been described, such a system that was already implemented (pre-slow-apocalypse) would be in effect. It would already be doing the action of preserving knowledge.

  • 1
    $\begingroup$ This was my first answer on worldbuilding. I found the OP question interesting and thought up a solution that could be implemented today. $\endgroup$
    – voxelv
    Dec 22, 2016 at 23:16
  • $\begingroup$ Microprocessors, or more generally transistors do age. Even when not operated, the doting agents slowly diffuse/dislocate in the silicon layer, altering the electrical properties. This is still a field of ongoing research, and there is precariously little hard data of actual product lifetimes. In the 80's i've heard claims that a contemporary CPU should last ~200 years, but since then structure size has been decreased a whole lot, probably worsening the life expectancy proportionally. From my personal experience, laser diodes in DVD drives have high failure rates after few years of operation. $\endgroup$
    – Durandal
    Dec 23, 2016 at 17:43

Laser-Engraved Data-Cubes

If you don't need to edit the data afterwards, using lasers to engrave the data in an information-dense format inside an artificial-diamond or Silica-Glass block would probably work well long-term.

This has the advantage that you can technically read it (painstakingly slowly) using an optical microscope and transcribing it onto a contemporary medium. Or you can place the block into a purpose-built reader and access it more quickly.

I imagine a 12cm cube of artificial diamond (2500 carats if you're curious) being able to contain quite a substantial amount of data in a binary or bitmap format.

It wouldn't be indestructible (diamond is generally pretty brittle and can be easily crushed to powder with a hammer-blow) but a large chunk like this affords quite a lot of robustness.

Surface damage could be buffed off, even chips knocked out of it wouldn't be enough to ruin it if the actual data is stored deeply enough inside the cube.

It would probably be sensible to make a lot of cubes for different subjects, as well as creating duplicates for redundancy.

For human-readable purposes, the individual cubes could be given an engraved serial-number and/or title text in the margin where it can be read.

In fact, there are people working on something similar and the technology holds some serious promise for long-term data storage, including the kind of extreme scenario you describe.


You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .