How to store data for a billion years | Hacker News Reader

How to store data for a billion years

economist.com

29 pointsrobg32 comments

> To switch spontaneously from a “1” to a “0” would entail the particle moving some 200 nanometres along the tube using thermal energy. At room temperature, the odds of that happening are once in a billion years.

So, if you stored a GiB of data, that puts the lifetime of that data at ... 1 year? That honestly doesn't seem very good!

That's pretty much what I was thinking, reading this. First off, a frequency measurement -- "once in a billion years", 3.16887646 × 10^-17 hertz (thank you, Google calculator) can't really be used as a probability, as they are trying to do.

Then, even if you assign that nonsense statement some sort of charitable reading, like "one bit has a 50% chance of spontaneously changing state over a one billion year period", and infer that every bit has a 1/10^9 chance of flipping every year, then the odds that 8×10^9 bits do not change at all over the course of a year are (1-10^-9)^(8×10^9), or about 0.034%.

Granted, it'll still take a long time for that whole gigabyte of data to be completely scrambled, but I think we'll still be using error correction and occasional refresh, even with a technology this reliable.

If you do a Raid 6 Array and manage to keep power to the storage devices you could fix the lost bits whenever they came up. Or you could just write to 10 drives at once. The odds might be that the whole does not change 0.034% of the time for one drive, but for 10 it would be > 99%.

It can be acceptable if you use CRC. You can lose a few bits and still be able to read your data.

Most media today are unreliable. I don't know what the numbers are for CDs, HD, or floppy disks, but the probability of spontaneous bit switching is supposedly higher.

To give the Economist their due, the real title is "Not Forgotten" and they make no claims about billion year data storage, "Theoretical studies suggest that the system should retain information for a long time." Has the OP been spending too much time over at slashdot? ;-)

The byline to the article is "How to store data for a billion years"

Apologies. I missed the byline. The writer may be restrained, but the editors/headline writers need a whack on the sides of their heads.

Only if the chance of this happening to each bit is independent, and also memoryless. Either or both of these might not be the case; I don't know.

You also need to build a machine that is able to read it in one billion years time. Anyone in the room that is still able to read a laserdisc? Didn't think so.

So, storing the information is one thing - being able to read it is a completely different beast. If you would really want to store something for a few thousand years, you're probably better off chiseling it in stone. Then you only have to make sure people (or aliens?) can read our letters and understand our words.

Well, so long as you clearly mark something as being a store of data, a sufficiently advanced civilization would be able to figure things out.

You could even place some dummy data before the main store, with redundant storage in another format - if you read ("decode") the data and it matches (like you suggest) something carved in a block of stone surrounding it, you know you have the process down, and you can go on and decode the main store.

Agreed. You would probably have to start by explaining that there's important data, and that it is stored at a nanoscale. Then you would have to proceed to explain how they might get access to that data.

But that's not the whole of it. Once they can read the 0's and 1's, you need to tell them how to interpret them. So you need to provide (at least) an ASCII table in some extremely durable material.

The rest of the information (on how to interpret the data) can then be stored in the first mega/gigabytes of the binary data. Plain text ASCII only, ofcourse. No pictures, audio, or video. You could only include such data once you've told them how to build a computer and all necessary software.

I think you're better off storing a computer and a self-sufficient powersource in a vacuum, virtually indestructible structure. Has the added benefit that you can make cool movies about it :)

>The rest of the information (on how to interpret the data) can then be stored in the first mega/gigabytes of the binary data. Plain text ASCII only, ofcourse. No pictures, audio, or video. You could only include such data once you've told them how to build a computer and all necessary software.

That's OK. At these data density scales, a small room full of 10"x10" tablets would provide space for a LOT.

I'd imagine you'd encode media in some simple lossless RLE format which you would then describe.

Imagine being an architect in the year 27,000 coming across this stuff. I'm sure the value would be immeasurable, especially the video and audio! Can you imagine having HD "documentaries" from Ancient Egypt or Greece? How cool would that be!

As you can see, this stuff fascinates me.

Now where to find those nanotubes that previous civilizations left behind for us?

I think you just figured out what those easter island statues are for.

Not to be a dick or anything, but we know precisely what the moai on Easter Island were for. They were religious figures of the Easter Islanders. They grew increasing large and numerous as the tribes of Easter Island began warring with one another, in an attempt to display the "greatness" of each tribes respective leaders. It is in part due to the obsession with symbolic greatness and the construction of the moai that led to the eventual downfall of the Easter Island culture, as can be seen by the eventual grass-roots moai destruction activity during the proper collapse of the island's society in the 1700s and 1800s. The few Easter Islanders that survived the end of their civilization described this back when they were first contacted.

Easter Island isn't some big mystery, unless you don't read history books.

</pedantry>

Whats pedantry to some is quite an interesting aside for others. Thank you, and keep it up :-)

I recall reading a science fiction novel where there were some proposals to store data for a long time.

Assuming sufficiently advanced technology, you could turn a seismically stable rocky planet into data storage by encoding data with valleys and mountains.

Simply encode your data into the the digits of pi and then create your own universe with this value - that's what I did!

I wonder if you could stored data for that long by encoding it into some junk DNA?

Junk DNA actually changes a whole lot because there's little or no evolutionary downside to mutations there.

That's a good point. I had assumed that since There's no evolutionary pressure it wouldn't be forced into changing as much as active DNA is. You could encode it to multiple areas that must be checked against each other.

That idea has received a number of scifi treatments. One example:

http://www.fourmilab.ch/documents/sftriple/gpic.html

Now we only need some data that somebody in a billion years wants to read.

This reminds me of the story of the traveller who was required to take all human knowledge to a distant planet carrying only one steel rod.

His solution?

Take everything ever written and convert in to binary. String it all together into a very very long string. Put a decimal point in front of this string. Then put a notch on the steel rod exactly that decimal's distance from the bottom.

I imagine they'd need some pretty precise measuring equipment on the other planet.

He'd need two marks. The other one is to tell you how to interpret the first, that is, the definition of 1.0.

Also, unless he was using a very long rod, he'd have to mark somewhere in the middle of an atom.

I think that issues like this are part of why length is now defined in terms of wavelengths of light. And even then, the number of digits that we use in the definition is relatively small.

and I thought platinum pins for 1s and holes for 0s melted into solid granite bed is the preferred long time storage!

Much better for very long term storage. The method described is actually for extremely dense storage; the time period is pretty much bogus journalistic stuff.

basically a nanotube abacus... very cool..

This is fascinating stuff-I believe if the human race can stop from utterly destroying itself we're on the brink of coming up with some truly spectacular advances in the next 50 years.

We are less likely to destroy ourself now than at any other point of time in history.

Why do you say that? After humans spread across the globe in hunter-gatherer societies, we were pretty unlikely to destroy ourselves.

Today we could easily destroy ourselves with nuclear war. It may not be likely but it's more likely than in prehistory.

Life in those days was very precarious. One bad drought and everyone would have been dead.

A nuclear war is not likely to kill everyone. Plus a nuclear war is not likely these days anyway.

The chance of a nuclear war today is far far lower than some disease in the middle ages, or a drought in pre-history.

Everyone focuses on nuclear weapons, but reality is they are very very unlikely to ever be used, plus illness is a much more severe danger. And the good thing about today is that we are much better prepared to handle a disease than at any other time in history.

And that's why we are less likely to destroy ourself now than at any other time.

Reply on news.ycombinator.com