Nanofiche: Small Storage, for Forever
archmission.org
archmission.org
In their case they end up having a "Museum of Humanity" that exists in man-made caves in Pluto specifically meant to merely have information carved into the walls as all digital and other physical formats weren't deemed sufficient.
I suspect these data formats are similarly not sufficient.
Digital ECC adds an extra layer, but analog text is still pretty good.
Nice example!
> ... worked for 1984.
Another nice example, but ouch, that hurt.
The things in 1984 are the natural outcome of a totalitarian authority seeking control. Orwell simply made his predictions on a phenomenon that existed.
I mean, I guess you don't need a computer to read this, but you sure do need a good microscope.
I guess I'm not really entirely sure of the point. If you're really trying to preserve basic knowledge to restart civilization, I'd think you'd want libraries of full-size readable books spread across the world. Not something that requires you to first build a microscope.
While if you're trying to store as much historical and technological data as possible, it seems like you'd want to encode it digitally with error correction codes, not as optically readable text. The encoding/decoding instructions can easily fit on a single cover page, optically encoded.
Can't read what no longer exists, by any means.
Paper disintegrates without maintenance. Tape disintegrates. Hard drives lose magnetization. SSDs lose their charge states. Stone erodes. Several of these won't survive fire, water, magnetic fields, or radiation.
Also, I don't see why you couldn't store digital data on this, for example as pits and lands, or base 64 text, or some other scheme, with error correction of your choice. (But if it's as durable as they say, it's not as necessary.) But, that requires additional effort to decode.
I enjoy research into highly stable means of storing data. I think Microsoft Research was developing a technique of storing data in a 3D arrangement in glass. Very fascinating.
Computers, peripherals, microscopes, and any other devices needed to access some incredibly durable archival format will also degrade. The information needed to produce new ones will need to be archived, and that material will degrade and need archiving.
Do you know how much information is necessary to describe all of the details relevant to producing an extremely good microscope from scratch? It's not just the parts and materials, it's the entirety of the supply chain: the mining, refining, electronics, power production, and everything else. All of that information needs to be accessibly preserved.
So it begs the question, is it really easier to make the prerequisite for the information's use be a working precision instrument or is it easier to just actively maintain lots of copies of less durable materials like books?
If the sky falls tomorrow, how many years is some archive really useful for in any format? Most of a sturdy, well-designed library of books will probably survive for a decade. A good percentage will probably survive a century. On the other hand, a high precision microscope probably won't "just work" after a decade or two of sitting around disused in unideal conditions. A working power supply would probably last even less time.
We must accept that there's some ongoing upkeep of an archive regardless of how durable its contents are. And while a book isn't as durable as some quartz or other fancy material, you can pretty easily make books to replace old ones even in pretty miserable conditions (we did it for centuries, albeit at great cost). There's a remarkably high cost to making sure there's a working microscope around, and keeping the information and resources necessary to make new ones and repair existing ones available.
The whole point is that the operation persists for longer than a human lifespan. If we have to rediscover 100 years of making lasers and rockets, who cares? If we have to rediscover 2000 years of agriculture and medicine, that's a problem.
But if you want to be doubly sure, include instructions for making a hand lens at 1x, and instructions for making a microscope at 5x.
> We must accept that there's some ongoing upkeep of an archive regardless of how durable its contents are.
That's a good point too. Archiving is a living process.
I have books on my bookshelf that are older than 100 years old. I don't think 100 years is even that much to ask, seeing as i know who William Shakesman and Chaucer are.
I'm in a unique position, though - my music is on an SD card traveling in space right now, thanks to some attentive friends who petitioned to get me on there. Or so i'm told. How would i know, my spaceship is in the shop.
And, in practice, it doesn't matter whether the storage medium can't be read because it's damaged or because the machine to read it doesn't exist anymore. Either you can read it or you can't.
It's not a prerequisite, you can have a smaller batch of knowledge on how to build the tools to access all the knowledge in various other more accessible systems.
Besides, you presuppose that just because there is no machine, there is also no knowledge re. how to build a machine and you only have twigs and leaves. That's a false universal assumption, you could just as well be in a situation where you don't have the machine, but you have tools and resources to build it. This is yet another difference to the "the medium is broken and can't be used" situation
a-la Superman ?
But if this is trying to entice libraries with "replace your microfiche with this easier to maintain format, and do something for history while you are at it" then you need to store scans of newspapers, magazines, etc. Which probably means storing PNGs or similar. At that point, I'm not sure the capacity improvement over "analog" optically readable images is really worth the much more complex decoding setup.
Both for the library use case and for the "a civilization that might not speak our language wants to decode this" use case the complexity of the decoding setup is a major factor.
(Obviously things like that are multi-causal, but I had never heard that idea before, that the amazing porcelain of China kind of held back the development of glass and all its many, many applications.)
It makes a lot of sense to make the most useful information for apocalypse survivors the most accessible information. Just make books about that information and keep them maintained. Then there's at least incentive to keep the books protected and maintained.
Edit: it sounds like you'd probably need a 1000x to 10000x microscope to really read the nanofiche described in this article. So a potter's wheel is probably not sufficient. Plus, you'd need the materials and tools for working with glass, which might not even be feasible in an apocalypse.
The tech that survives is probably an odd amalgamation, but chemistry is basically the thing you really want for society - antibiotics, explosives, and fuel.
That is if you have a thousand year gap in your archival material warehouse maintenance the data is lost. If you have a thousand year gap in your ability to make microscopes the data will wait around patiently.
Except you don’t need a “facility”. A lot of lot of knowledge fits on a bank card size nanofiche. You make ten thousand of them, and people put them away as best as they can. Some will melt in a fire, or get blasted apart, some will be burried and forgotten forever. But some will be more lucky, by just chance they will end up in an atick or crawlspace where nobody disturbs them.
Not counting that we can hide some of them in places where they are unlikely to be disturbed, such as place a few on the moon or place a copy on every single geo and lagrange point destined satelite. Heck, burry a few on a glacier and they will naturally re-appear in a few thousand years.
The hard problem is meting through the information. Fast, contextual search is probably the most important development we have today and you can’t really replace it with anything but tons of compute and digitization. Big ass physical libraries don’t solve this either.
> Wikipedia printed on a nickel book with each leaflet holding 8,000 pages
It's tiny...
It doesn’t have to be an instanteous thing. Just think of how much time schollars spend on deciphering papyrus fragments. If we would have a nanofiche of the contents of the library of Alexandria multiple departments would spend multiple lifetimes reading and contextualising it.
But that is the beauty of a microscope. It is not some “unrecognizable tech”. Seeing tiny things in detail is a basic skill many new civilisations can re-discover on their own. It doesn’t matter if it is our human descendants, smart octopuses or the silurians. They will be interested in seeing tiny things and thus they will construct their equivalent of microscopes.
The important thing is sign-posting. How will they know where to look?
Of course you'd want to do both (early steps, and advanced knowledge).
And there will be various kinds of low-hanging fruit, to propel newly-industrialized society into the space faring internet age. Skipping resource-wasting, fossil burning stages entirely.
As for culture (novels, music, games, software, movies, etc): who cares, whatever makes it through the bottleneck in whatever way, shape or form.
It's not an either/or though. Optimistically we'll be able to store all our collective knowledge in all accessible formats, digital / internet accessible, libraries, etc.
But nobody can see the future. Having this as a backup sounds great. Archive all of e.g. Wikipedia, put it in a time capsule along with a microscope and/or readable instructions on how to make one, make thousands of them and put them in landmarks, caves, monuments, dumb sattelites that are designed to land on ground in 10.000 years, the moon, around the sun with a homing signal (e.g. reflective surface since electronics won't work after 1000 years), etc.
A library can catch fire, the earth may get glassed in WW3, but surely some of these would then survive for future generations or alien archeologists in the millennia to come.
"Sir, we've set up our 3 kilometre array of precisely calibrated space telescopes, capable of detecting the slightest wobble in the deep space radio frequencies, and automatically translating all possible alien languages!"
Aliens: No bueno. You'll actually need to find the 1 millimetre disc of quantum-lithographed Nickel buried on one of our physical moons, 1000 light years from your location. There, you'll find our entire history and knowledge! Enjoy! ;)
Imagine a stretch of desert extending past the horizon in all directions, each grain of sand containing a wikipedia-worth of information, mostly redundant. Hopefully useful, and not maddening like the monolith from 2001.
https://www.microsoft.com/en-us/research/project/project-sil...
Long-term archival storage is going to be an important industry as our data volumes grow.
Unless you care about reading your data, as then it won't be important. Let me know when tape starts taking more of the global storage market share. It's a super mature, functional, and far higher performance storage media, and it's total percentage of global bytes shipped dwindle.
It is sad there is no removable media other than maybe these few TB portable ssds that can be used for archival purposes. I _still_ have my commodore 64 datasette tapes as well as floppies from the start of the pc era(for me).
Even for the prosumer there were iomega drives. Now, one is supposed to use (and trust in) the cloud. Bonkers.
Exactly! Or where is the affordable 1-5 TB Tape?!
And worst of all SSD is not reliable. And apart from an over supply market where prices have fallen to record low we are at right now. NAND cost wont be dropping much in the near future.
Apart from big Enterprise, most smaller shops dont know and cant afford to do storage / backups.
I know the Long Now foundation has the Rosetta project but that doesn't seem very live. It would seem straightforward to use technology like this to create 10,000 copies of something that could be widely distributed.
Don’t get me wrong: love the goal. Practically tho it’s basically just nerd hobby projects
Could this be done in one step?
The LongNow used it for some random codex & then seems to have dropped it. Last I saw it was mostly used by one single jeweler, bizarrely.
People nowadays are struggling to find devices (and software drivers) to read old formats/mediums from a few decades back, imagine how difficult it’s going to be to find a working nanofiche device and support for it under Windows 9000 in the year 9000AD.
I like the idea, the nickel based storage medium, the improvements in capacity and all the other advantages listed, but I don’t like the proprietary technology.
A nanofice "reader" is literally just a microscope, isn't it?. Probably a very good microscope, which may or may not be difficult, given the technology of the time, but generally thought to be understandable. You'd store a bunch of this 'nanofiche' in a place, and have some diagrams/example documents with small text getting smaller and smaller and smaller, and the curious people/aliens would (hopefully) understand the idea, know enough about optics to magnify, and boom, they've got the data.
I could imagine they would not want to make a single page too small (makes the readers more complex) so perhaps they draw the low-res images a bit bigger than needed.
It's hard to take someone seriously when they claim to be acting on behalf of all humanity then immediately does something as pretentious as this. Probably made for the same kind of elitist types who like to spit on you as they try to demonstrate how they can pronounce "LaTeX" correctly.
All this depends on the font used, and what kind of information is stored. But using lorem ipsum, and Verdana, and only uppercase characters, spaces, and periods, I found that you can fit about 49,500 characters in a letter format page at "150 dpi" [0]
This is 396 million characters in a 2 x 2 cm sheet. The Latin ASCII alphabet has 26 letters, but in my experiment, there were commas and periods. Though at the resolution I used, commas and periods would be difficult do differentiate if there was almost any degradation or printing error. So let's say it's a 27 character set. Maybe we can stick a few more numbers or symbols in there to get to 32, although some numbers like 1, 0, and 5 look like letters, so not all would be usable at this resolution.
If each character can represent 32 states, that is equivalent to 5 bits per character. 5 x 396 million is 1.98 short billion bits. That is 247.5 million bytes. Or 247.5 megabytes. This is in 4 square centimeters, so the data density is about 61.9 megabytes per square centimeter.
If we give up human readability and the text would be regular language (according to Zipf's Law), we could expect up to 0.2 compression ratios. Or 310 megabytes per square centimeter. On the other hand, we could then use quite dense data storage methods like barcodes. But these calculations are a bit pointless because this is meant to be read without a computer.
A DVD has a readable surface area of about 89.2 square centimeters on each side (some DVDs are multi-sided) and 4.7 to 8.5 gigabytes of storage in that space. So it stores between 54 and 97.6 megabytes per square centimeter.
For storing plain text, it might be on the same order of magnitude in terms of space efficiency as a DVD, but it is human-readable. If the data is not human readable, then it could be at least 5 times more efficient than a DVD. Which is closer to a Blu-ray disk.
Back-of-the-napkin, might contain errors, many assumptions made. The goal is to understand the order of magnitude with some errors. In short - it looks like it's not 100x more or less efficient than writing a plain text file to a DVD.
P.S. I chose Verdana as it was a quite legible sans-serif font. I then created a texture of random black and white pixels, and applied it at 25% opacity to simulate a significant degree of printing error at the supposed resolution they can print at. The font at the chosen size remained just about legible with that kind of interference [1]. The exception is periods and commas, which are difficult to tell apart. That is why I considered them one character before. Making the font even 10% smaller would make it really difficult to read. This is all not grounded in any technique or proper estimation method, I just wanted to show why it seems like going smaller might be difficult.
[0] https://i.imgur.com/Pi8acSk.png [1] https://i.imgur.com/stFkXQy.png
1)Demonstrably no novelty whatsoever. "My device does what they do just a bit smaller" isn't novelty unless the processes themselves are some kind of breakthrough which these clearly aren't.
2)Not being obvious is one of the conditions for a patent is the US[1]. It seems you could make a strong argument for this being obvious in the sense that
a)it is. "We have a thing that stores information by making it small. We want to store more information... I know let's make it smaller." is definitely an obvious chain of reasoning
b)to make it more obvious they even helped your argument by naming their thing after the one that's a bit bigger and just changed the name to the next engineering prefix down in scale.
3)Very clearly undermines their whole facade about doing something for the benefit of humanity.
Fuck any proprietary format trying to own the ability to recover history.
we go through cycles of anti-intellectualism as a species and the anti-intellectuals will always find a way to attack knowledge regardless of technology used to preserve it. if you carve stones, they will break them. if you write books they will burn them. if you make an internet they will firewall it or buy it or sue it. if you put plates of nickel on the moon, they will figure out how to go there, dig it up, and destroy it too. but most of all what they do is attack the idea of the intellectual as valuable to society.
Look at how we treat Aaron Swartz and Alexandra Elbakyan. We treat them like criminals, while people who jail and murder intellectuals we call leaders and presidents. We live in an anti-intellectual era right now.
Not that it is not important to try. but there are some problems we cant solve by tech alone. We have to figure out what is wrong with us, as a species, that we constantly try to destroy our own knowledge.
As you said culture is probably the best method of preservation. But still fallible, and this may be a good thing!
How do we expect someone will be curious enough or have right gear to read these sheets?
~.1 micron. Pretty small feature size. Typical letters might be 30 by 15 dots in a 10pt font with their 150DPI assumption. But it's not -that- exotic.
A cheap compound optical microscope might do .4 micron, resulting in letters being 8 times and 4 times the spatial resolution. This is in the range where things are slightly annoying in fuzziness but still easily legible.
My crappy LCD microscope for my classroom cost about $50 and has a resolution of roughly 1 micron. It could probably read, with difficulty, 12pt text.
i'm more worried about how scratch or defect resistant it is, given the small feature size. if you have to store it in some kind of case to protect against scratches, then how long does the case last? do you have to do this in ultra clean room conditions, to prevent fungus etc from being encased with it?
Doesn't seem tremendously reassuring.
https://www.space.com/fungus-in-space-long-duration-astronau...
> In 1988, astronauts aboard the now-retired Russian space station Mir realized that something had blanketed one of their windows — from the outside. [...] This fungi had managed to adapt to the space environment, and so well that it not only survived but thrived on windows, control panels, air conditioners and cable insulators. It even contaminated the crew's precious food and water supply.
Not a one-time event.
So unless the nanofiches are somehow stuck to the heat exchanger of a secret lunar base or similar, I'd assume that they would go biologically inert (or at least dormant) very quickly.
I’m not sure I would’ve marketed it exactly this way. But this is because I spent almost a decade in the Danish public sector and I know how revolutionary long term storage will be for public recording. Right now we spent a very high amount of resources on the maintenance of data records that are rarely accessed but are important to public history, accountability and many social sciences. If you could sell those recording organisations a medium that didn’t need near constant maintaining it would free up a lot of resources. Still, maybe it’s a good choice for a company mission considering my earlier points as this tech will essentially sell itself. If it works as well as it’s sold in this article.
In the content of hacker news, however, why wouldn’t you expect this to be meant to make its investors money?
1)Maybe I'm wrong but I don't think their IP has any kind of protectable moat. I really can't see that patent surviving any sort of challenge. So if tehy figure out a business model that works it will get cloned cheaper.
2)It's an actuarial fact that people overpay for insurance relative to expectation but here they are essentially selling a partial mitigation for a very high severity very low probability event[1]. So the expectation is very low and is highly subject to model calibration error. Potential customers are not going to value this well and I don't see people paying them enough to stay in business.
I could easily be wrong about both of those things.
I know about selling the mission "Making the world a better place with extensible mobile, social, local middleware APIs" etc as I think the TV show "Silicon Valley" parodied it. Just in this case I dont' really see either the mission or the commercial reality making much sense.
It's interesting that the board of advisors has Ben "I'm gonna bring the mammoth back to life" Lamm and David "Make the statue of Liberty disappear" Copperfield on it. So they've got Barnum and Bailey going for them at least.
[1] "I'm selling you a backup that would survive even if humanity gets wiped out" isn't a business proposition. Noone cares about what their backups do under those circumstances so they won't pay for it.
This goes back to 1995 or so: https://rosettaproject.org/