I'm designing an alien society with a low population density. I intend for them to have 1 book per topic per place; books update annually as new information arises. I call them books, but they can even be large epigraphs and are stored in optimal conditions (e.g., in a cool, dry, and dark room where they are made and treated with care and cured of mould and insects regularly with little air but many preserving chemicals). If your method requires conditions that aren't described here, tell me.

All books are on/in:

  • Hard drives and the Internet, which are rewritten frequently to prevent data corruption and loss
  • 1 national library
  • Archives (e.g., underground in bunkers)
  • Possibly off-planet on extraterrestrial missions as necessary, and
  • Citizens' houses.

However, despite updates, if I still want them to be preserved forever or for as long as possible, what substance do I use? Note that I suspend disbelief within reason as long as a substance may theoretically be inscribed upon and fit into a single planet and preferably a single city, so aluminium, graphene, sapphires, etc. are options. My reasoning is that my species is focused on scientific advancement and probably won't want to be tasked with repairing broken books after each decade.

Note on Physics: I don't care about Heat Death. I just want the best book materials within this universe. If that means replacing them every few centuries, so be it, but I want replacement to happen as little as possible.

Edit: Upon prompting, I have changed the word "epigraphies" to "epigraphs," which I mention to avoid confusion. Someone else has changed the wording and formatting a little bit before I could find the time to do it, which I would like to thank them for. If any of this impacts your answer, bear it in mind.

  • $\begingroup$ Welcome to Worldbuilding.SE NimNim. Please check out our tour and help center. I notice your tags were just science and every major type of science, which I assume you meant to indicate the content of the books. But the content doesn't matter because your question is about preservation of the books. I agree with Renan's change of science to science-based but please read up on that tag to make sure it's what you want (it places restrictions on the answers). Make sure I got it right too. $\endgroup$
    – Cyn
    Jul 7, 2019 at 2:21
  • $\begingroup$ What you need is a Distributed Version Controlled Database Management System. This is not a long way from what github.com provides. You might also investigate arxiv.com a repository of scientific papers which also maintains different revisions of papers. You need to investigate how systems like these (and e.g. financial records) are maintained in computing. $\endgroup$ Jul 7, 2019 at 2:29
  • $\begingroup$ Frederik Pohl's Heechee developed technology rather along these lines. Humans, who did not recognize them as data storage, called them "prayer fans" and sold them as souvenirs. $\endgroup$
    – user535733
    Jul 7, 2019 at 3:16
  • 1
    $\begingroup$ I suspect the biggest problem is that even if your books survive a million years, nobody will be able to decode them; languages drift over time. Mere thousands of years are already a massive gap that we weren't able to bridge for some languages, and thus the knowledge written down was lost. $\endgroup$
    – Luaan
    Jul 8, 2019 at 12:28
  • 1
    $\begingroup$ One wonders how on earth a question is marked as "unclear what you're asking" when twelve people (modulo some duplicate answers) have had no problem interpreting it and producing answers fitting a common theme, and many more people have decided that those said answers are good :-/ $\endgroup$ Jul 8, 2019 at 20:12

11 Answers 11


On a living planet, nothing lasts forever. Especially if you're limiting the number of iterations of each book to one. One book per topic (e.g. Shakespeare) per place (is that household / city / county / nation / planet?) is asking for disaster, and even more disaster as you move from left to right along that scheme!

That said, there are several time tested possibilities:

Clay Tablets

The "Assyrian National Library", sponsored by Emperor Ashurbanipal at Nineveh, is the oldest surviving royal library known, dating to 700BC or thereabouts. Clay is easy to write in and once fired, the tablets are sturdy. They're a pain in the backside to deal with for texts of any great length or complexity. Can last thousands of years if well cared for, or if abandoned and forgotten for millennia.


Egyptians carved literature into stone. Like clay it's durable, but is not indestructible. Like clay, it's a pain to maintain lengthy texts. Can last myriades to millions of years if protected from erosion

Papyrus, Paper & Vellum

Well known to ancients & moderns alike. Easy to make & convenient to write on. Durable and long lasting. If you're people are repairing books every decade, then I'd say they are either overly abusive towards books or else don't know to bind books. A book printed on good rag paper and well bound can last several centuries without needing more than minor repairs. When well maintained, books of this sort can last many centuries: expect somewhere between 13 and 20 centuries. All these materials will eventually decay, even in the best of conservatories.


This is probably the most durable option. Books have been written on gold pages, the oldest of which are in the region of 2500+ years old. Copper / bronze, stainless steel, gold, platinum: any of these should last indefinitely under stable atmospheric conditions.

To Digitise or Not to Digitise?

Even the best digital media will find it difficult to compete with well curated books. Digital media are also heavily reliant on external factors such as electricity, computers, operating systems, stable cultures & societies just in order to be accessed & processed. Whether it's magnetic or battery supported or flash memory, digital media are not a good idea for truly long term archival of a culture's most important data.


I'd recommend that the culture's most precious data --- scripture, literature, mythology, folklore, poetry and the like --- be preserved on metal plates. And also on high quality paper books bound in solid & protective hard covers.

All rapidly changing data --- scientific & technical developments, rapid advances in medicine, changes in law --- should be committed to okay quality paper.

Ephemeral data --- accounts, receipts, bills, notes and so forth --- these don't need to be archived at all and can be kept on cheap paper.

  • 2
    $\begingroup$ Aligns with my thoughts on the matter. +1 $\endgroup$ Jul 7, 2019 at 2:04
  • 3
    $\begingroup$ I would argue that stone is probably more durable than most metals. If you're going to use a metal, make sure you use one which doesn't corrode in your atmosphere. Gold and Platinum are both good, if expensive, choices. Depending on your use case, you could laser etch into metal or stone and use it like microfiche. $\endgroup$
    – Turksarama
    Jul 7, 2019 at 22:12
  • 2
    $\begingroup$ Digital is not a material, and should not be referenced as such, it's an encoding scheme that helps protect against degradation, and combining with error correcting codes is extremely robust to damage, digital will always be superior to schemes not using it, No one is stopping you from putting a QR code on paper, destroying 30% of the data, using a ruler and a pen and paper you can recover the entire text because of the ECCs. $\endgroup$ Jul 7, 2019 at 22:38
  • 1
    $\begingroup$ @Turksarama -- Excellent point for specifying atmosphere. For purposes of this answer, I chose an atmosphere much like Earth's. YMMV. As for the durability of stone, I agree stone itself is quite durable. The problem is carvings into stone are often not deep enough to endure; and would have to be very large and very deep for any amount of detail to survive anything like "forever"! A stroll through an old cemetery demonstrates that even carvings in hard stone wear out relatively quick. $\endgroup$
    – elemtilas
    Jul 7, 2019 at 23:14
  • 1
    $\begingroup$ Text is already digital, it just uses an encoding that's optimized for decoding with convolutional neural network recognizers instead of traditional error-correcting codes, just because that's what our ancestors had available. $\endgroup$ Jul 8, 2019 at 6:56

Fused Quartz Etched by Femtosecond Laser

Some articles on this technique here, here and here, and a wikipedia article with more references here. According to that first article,

The current data-writing system is not much different from that found in CD or DVD drives. Ultrashort laser pulses with a wavelength of 1030 nm are focused inside a spinning glass disc and the position, power and polarization of each pulse are simultaneously modulated depending on the encoded information – leaving a trace of pits with different optical characteristics. Reading the data is more complicated because it requires a microscope-based birefringence measurement system, but we are now working on how to solve this problem.

The original paper says that they estimate how long the data will last by looking observed decay rate of the nanovoids (the 'pits' made by the laser as mentioned above) at "several annealing temperatures in the range from 1173 to 1373 K", and then using the Arrhenius equation to extrapolate the decay rate at other temperatures. In fig. 4 they present the following chart showing the "thermally activated decay time" $\tau$ (which they mention is equal to $1 / k$, where $k$ is the decay rate in the Arrhenius equation) as a function of the temperature $T$:

enter image description here

So, though one would have to preserve the fused quartz records in a place where they will be extremely well-protected from shattering (as fused quartz is a type of glass), the time that would pass before the information would degrade due to ordinary thermal decay is extremely long--longer than the current age of the universe (13.8 billion years) at a temperature of 462 K (189 C) or less, and $3 * 10^{20}$ years at a room temperature of 303 K (30 C).

More generally, ever-increasing number of backups are needed for arbitrarily long timespans

Ultimately if you are concerned about your civilization preserving the information indefinitely on cosmological timescales (as suggested by your comment about assuming they can avoid heat death and proton decay), you will want the civilization to periodically make new backups and store them in different locations throughout the universe, so that the probability that all records of some information are destroyed is continually decreasing over time. If the probability of all records of some information getting destroyed isn't decreasing this way, if you wait long enough it becomes a virtual certainty you'll lose that information. Say in a given million-year timespan the probability is $q$ that the civilization loses some item of information due to all records of it getting destroyed, so the probability the information is preserved in that timespan is $(1 - q)$. Then naturally if the probability is same in the next million years the total probability the information will be preserved for 2 million years will be $(1 - q)*(1 - q) = (1 - q)^2$, if the probability remains constant for 3 million years the probability the information is preserved in that time is $(1 - q)^3$, and so forth. No matter how close $(1 - q)$ is to 1, there's going to be some sufficiently large exponent $N$ such that $(1 - q)^N$ becomes arbitrarily small.

On the other hand, suppose the probability the information is preserved in the first million years is still $(1 - q)$ but the probability it's preserved in the next million year span is $(1 - q^2)$ and the probability it's preserved in the next million year span after that is $(1 - q^3)$ and so forth. So here the probability the information is preserved for 3 million years is $(1 - q^1)*(1 - q^2)*(1 - q^3) = \prod_{k=1}^{3} (1 - q^k)$, using Pi notation for products akin to Sigma notation for sums. Then if that pattern continues indefinitely the probability the information is preserved approaches a nonzero limit $\prod_{k=1}^{\infty} (1 - q^k)$, which according to this mathematica page is given by the Euler function $\phi (q)$, and the page also shows a graph of its value for different values of $q$. So this limit can be thought of as the probability the information is preserved forever, assuming a universe where a civilization surviving forever is physically possible (I talked about that question in this answer), and where they are able to create an ever-decreasing probability of losing all copies of some record by the method of ever-increasing numbers of backups.

  • 2
    $\begingroup$ This is the correct answer, and the only one seriously being considered by scientists. $\endgroup$
    – forest
    Jul 8, 2019 at 7:04
  • 1
    $\begingroup$ Just make sure you encase them in something that is a radiation barrier fission tracks will greatly increase your degradation rate. $\endgroup$
    – John
    Jul 8, 2019 at 13:51

Plain old ink on vellum scrolls.

You want books that can be added to and that don't need replacing very often. They don't have to "live forever" as your title implies, because you say "my species is focussed upon scientific advancement and probably won't want to be tasked with repairing broken books after each decade." So repairing every couple hundred years and replacing every (half) millennium should work.

enter image description here

This is a closeup of the actual Torah scroll at my synagogue. It's written with a natural ink designed for preservation on deerskin vellum (deerskin is very unusual but allowed). The scroll was written by hand in the 1700's and used regularly by a synagogue in Europe until stolen by the Nazis (who murdered all the residents of the village) and put in a warehouse for decades.

It was donated to a nearby synagogue likely in the 1960's, then transferred to us. It lives in a cabinet that has no special environmental controls. The scroll consists of sections of vellum stitched together then rolled up on wooden rollers. It is tied shut with a ribbon then covered with a cloth and stored upright.

enter image description here

This scroll is handled regularly. Many synagogues pull out their scrolls and read them several times a week (they often have several scrolls they rotate), others less often. Any given scroll might be exposed to air and movement a couple dozen times a year. We use implements to touch the writing so our hands do not leave oil and dirt, and the scrolls are handled with care, but basically there isn't a lot of special treatment. We never wear gloves or anything like that.

There is some maintenance involved, but nothing major. The scrolls should be cleaned every few years. The covers get replaced now and then. I'm not sure if our wooden rollers are original (probably not) but every 100 years wouldn't be unreasonable for replacing them (your society could use a more durable material, like metal).

So this scroll which is around 250 years old is definitely fading and has a couple of stains and the vellum is eroding a bit on the edges. But it's still in use and is completely readable.

For a non-sacred scroll, go ahead and use a printer to apply the ink (there's nothing special about doing it by hand in terms of preservation). If you want to add to the book, just sew on a new panel (on the ends or even in the middle...panels do get resewn if needed for maintenance so this is very doable).

Will it last a millennium? Perhaps not. But half a millennium is very likely, baring disaster or poor caretaking. Pieces of scrolls almost 2000 years old are still readable, but going that long does deteriorate them, so I wouldn't go longer than 500 years (less is safer but perhaps not necessary).

So what do you need:

  • Basic reasonable care in avoiding finger touching, using a cover, keeping it dry, etc.
  • A light cleaning and check over every 5 years or so.
  • Replace the cover once or thrice every 100 years.
  • Consider replacing wooden rollers every 100-200 years, or use another material.
  • Sew in new panels as needed.
  • Reprint/write every 500 years or so.
  • 1
    $\begingroup$ If you have to reprint every 500 years, if we're talking about a very long-lived civilization (the parts about heat death and proton decay in the question suggest the OP is thinking in multibillion-year timescales), aren't copying errors likely to accumulate, whether you use hand-copying or something like scanning + OCR + reprinting? $\endgroup$
    – Hypnosifl
    Jul 7, 2019 at 17:59
  • 1
    $\begingroup$ @Hypnosifl There are cultural safeguards for that, but they work well for, say, Torah scrolls over a few thousand years. Not sure anything works longer than that. But OP says there are digital copies that are "Hard drives and the Internet, which are rewritten frequently to prevent data corruption and loss." So printed copies can be periodically checked against an original. It's also hard to imagine something in multi-billion year timescales that are stored in "Citizens' houses" or even an archival library. Nothing lasts that long. $\endgroup$
    – Cyn
    Jul 7, 2019 at 23:07
  • $\begingroup$ @Hypnosifl with a bit of thought, it should be possible to create some checksums that might be applied per paragraph, page or chapter (or whatever) that might help catch certain categories of change. This needn't even require a computer, though I'm not aware of this technique ever having been used historically. $\endgroup$ Jul 8, 2019 at 12:10
  • 1
    $\begingroup$ The UK still stores copies of its laws on vellum in the Tower of London. Why vellum? Because they know it lasts at least a thousand years, and modern paper hasn't been around long enough to have proven itself. $\endgroup$
    – workerjoe
    Jul 8, 2019 at 17:24
  • $\begingroup$ @workerjoe Maybe I underestimated the power of vellum. Thanks for that. $\endgroup$
    – Cyn
    Jul 8, 2019 at 18:29

To recycle one of my other answers, have a read up on the Rosetta Disk. This is an information storage system intended to survive at least 10000 years and still be consumably without seious technology.

Rosetta Disk

Basically, you use techniques derived from semiconductor fabrication to inscribe information onto a physically robust substrate that can then be read back via visible light or electron microscopy. You can aid this process with mechanical and electronic devices. Information density is limited, but its toughness and longevity is much higher than any electronic storage mechanism. Whenever you do an update, just mark your old copies as obsolete (and make sure your new copy is correctly versioned and dated so you don't read old stuff by accident!).

  • $\begingroup$ A single one of these is said to hold 13000 pages of documentation... $\endgroup$
    – Efialtes
    Jul 8, 2019 at 11:58
  • $\begingroup$ Just be careful not to scratch it! $\endgroup$ Jul 8, 2019 at 19:38
  • $\begingroup$ @Benjamin you're limited only by your ability to create hard substrates or plate etched substrates with transparent protective layers. The technique here, "Rosetta-HD" is used by de Beers to mark their diamonds, for example. $\endgroup$ Jul 8, 2019 at 20:08

Grey Goo

Sure it looks like a book. It feels like paper. But it is actually a colony of self-replicating, self-repairing, self-updating nanites. They are not sentient, and are designed in such a way that they can't run wild and consume the whole planet or awaken.

Any damage is repaired by repairing the individual drones, or disassembling and building new ones.

When I say self-updating, I simply mean the knowledge they contain and display. It maybe automatic, or may be through a specific process. It maybe that the nanites are able to translate words written on them in ink/pencil into their database, probably comparing and validating it.

  • 1
    $\begingroup$ a whole new take on e-paper. $\endgroup$
    – Jasen
    Jul 7, 2019 at 11:27
  • 1
    $\begingroup$ At least until a few millennia of cumulative errors leads to the world being converted into nothing but copies of Philip K Dick’s Autofac. Then you just feel a touch silly. Also like a book. $\endgroup$
    – Joe Bloggs
    Jul 7, 2019 at 19:29
  • 1
    $\begingroup$ @JoeBloggs there is no "cumulative error" in digital media, this is some big misconception about digital data that doesn't seem to disappear, the brilliance of digital data using error correcting codes is that it repairs it's errors en.wikipedia.org/wiki/Reed%E2%80%93Solomon_error_correction $\endgroup$ Jul 7, 2019 at 22:47
  • 1
    $\begingroup$ @DanielHill: Cumulative errors certainly can appear: it’s just that with a robust error correction code it’s vanishingly unlikely and can always be corrected post-hoc. Unless, of course, you’re talking about self replicating code distributed across trillions of processing nodes, where a freak one-in-a-trillion uncorrected set of bitflips that leads to a marginal uptick in replicator action is not only expected to happen once, when it does it’s going to replicate itself. $\endgroup$
    – Joe Bloggs
    Jul 7, 2019 at 23:54
  • 2
    $\begingroup$ Y’know, given that my initial comment was an attempt at light hearted humour, I’m not going to fight you on this one. $\endgroup$
    – Joe Bloggs
    Jul 8, 2019 at 8:19

Rock and Bone

We have good bone fossils from hundreds of millions of years ago, and other fossils are even older (up to almost half a billion years for some bacteria).

Your books are fossils encased in cubes made of bedrock. They are read through ultraprecise echography.

  • 1
    $\begingroup$ I sure didn't expect this answer.+1 Bizarre but makes perfect sense. $\endgroup$ Jul 7, 2019 at 2:03
  • $\begingroup$ I'm guessing "ecography" is a typo for "echography"? $\endgroup$ Jul 7, 2019 at 13:24
  • $\begingroup$ @TannerSwett you're right, thanks :) $\endgroup$ Jul 7, 2019 at 13:44

So I became quite interested in your question recently and I ran across this interesting wikipedia article:


It is about optical data storage, it uses femtosecond lasers to etch nanostructures into fused quartz crystal or fused silica glass. This techniques uses multiple layers etched within the crystal so that just a coin about an inch in diameter can hold 360 terabytes.

enter image description here

According to the article the crystal will last without decay for about 13.8 billion years, and is easily readable with a microscope and a polarizer. The fact that it is 5 dimensional means that if you view it from different angles different information would show up.


Alternatively, you could have a self-reproducing encoding e.g. put the information in otherwise unused blocks of DNA in a popular house pet.

There was also a novel (the title of which escapes me) in which an artificial intelligence backed itself up using steganography to encode data in elaborate and beautiful tattoos which it generated for enthusiastic human volunteers.


Not really an hard answer, more like some food for thought.

You are asking a technical question about material science but you need a different solution for your problem.

The main problem you will face is on how future generations will treat the writing, not if the writing material should persist times. We have perfectly preserved writings (3500BC) from the sumerians because nobody cared to loot or destroy them. On the other side, we have widespread destructions of historical artefacts (Worth reading: 2, 3, 4, 5, 6 and 7). Making an indestructible book by better writing material won't help as future brings better destruction tools.

I suggest using cheap, but durable materials, as printed ceramics. Be wary on digital formats, as those can be quite short-living. Perhaps in 100 years nobody understands .tiff or .doc anymore. Plaintext, as it is the easiest format, could work. Copy and distribute them over several, globaly distributed archives. Organize your archive in such a way that it will be maintained and future information ammended.


Store the data digitally in DNA.

DNA is incredibly dense from an information point-of-view, and lasts a very long time. In the natural world it has a half-life of 500-ish years, so with forward error correction and other preservation techniques could be expected to be 100% retrieved for at the least thousands of years. Much longer, if you allow for some degradation.

That's the cutting edge for human technology, available right now. Add in some advanced alien tech and you could easily imagine adding a few more orders of magnitude. How long is long enough? Seems like you could reach hundreds of millions of years without too much of a stretch.


Potshards are highly durable. A potshard is a fragment of pottery found in an archaeological excavation. In many cases an entire cup, bowl, amphora, etc. is found in pieces, crushed by the weight of material on top of it, and is glued back together and displayed in a museum.

So if a way to manufacture potshards (instead of entire pots) of the right size and shape to resist further breakage is developed, they can than be printed with data. Presumably a printing press can be developed with plates curved in the same way as the potshards and the pieces of type can be positioned in the plate similarly to regular printer procedures.

Possibly the press will print on the surface of the potshard with ink of some kind and it will then be glazed to protect the ink.

Or possibly the potshard will be still soft and the type will press into the soft clay and make depressions for the letters and numbers. Then the potshard will be fired to harden it and make the letter & number shaped depressions permanent. Possibly material of a highly contrasting color will be poured into the depressions to make them more visible.

Since all the potshards will manufactured with the same sizes and the same degree of curvature, they should be simple to store. And if there is a disaster they should still be legible millennia later if found by archaeologists.

Of course if the society is advanced enough to have digital storage methods, it will have more advanced methods of creating potshard like "pages" of writing. And may other possible materials, not just pottery, for the archival writing.

  • 4
    $\begingroup$ "if a way to manufacture potshards (instead of entire pots)". That's called putting clay tablets in a kiln. $\endgroup$
    – RonJohn
    Jul 7, 2019 at 18:55

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