Permanence of Data storage

[CaptainChewbacca]

I like both history and science fiction, which leads me to ponder what sort of things will be around for future archaeologists to study. How long does stored information last? If you leave a thumb drive in a dry place for a hundred years, will I still be able to get the data out? How about 1,000 years? How long do CD's last?

Basically, I'm just opening this thread to ask for any info those who are more computer-savvy may have about long-term data storage methods.

[Darth Wong]

CDs only last for about 15 years, according to their own manufacturers. Books last a few hundred years before disintegrating into powder. You can forget about magnetic storage lasting for centuries. Most things become degraded by environmental and bacterial action over time. If we want information to stick around for a really long time, we probably want to etch it onto gold and then put it on the Moon, or use the Flintstones' "carve it into stone" technique.

[Count Chocula]

IMO a thumb drive will last longer. Apart from electric shock, it's a more compact, durable storage medium, fairly well sealed, with convenient conductors for access by whatever data acquisition tools will exist in 100 years.

CDs and DVDs use a wafer-thin reflective substrate, with a pattern of burn marks, to store data. The platter is fairly brittle plastic, which can easily break or degrade over time. Plus, the shiny surface is wide open to physical damage.

My vote goes to the thumb drive. I've been storing my backup files on hard drives and SD cards for several years now, for the reasons stated above, and fully expect to be able to read them for the next few decades at least.

[Darth Wong]

Whooops, turns out that flash drives can potentially last for centuries, although the current generations of them won't.

http://gizmodo.com/5026149/japanese-sci ... s-of-years

That makes them potentially as long-lasting as paper books, although that's still not so long-lasting that some distant future archaeologist would be able to use them (especially since he won't know the protocols at all).

[Count Chocula]

In the future, every archaeologist will have in his tool kit a computer with every operating system and office, communications, drafting and art program ever written. With interfaces.

If only today were that day...

[Darth Wong]

In the future, every archaeologist will have in his tool kit a computer with every operating system and office, communications, drafting and art program ever written. With interfaces.

If only today were that day...

Who said they would be the same interfaces we use today? If I gave you a 1960s-era magnetic tape reel and told you to get me the data off it, how long would it take you?

[Ghost Rider]

In the future, every archaeologist will have in his tool kit a computer with every operating system and office, communications, drafting and art program ever written. With interfaces.

If only today were that day...

This has to be one of the dumber generalizations of what computers are. If I gave you a 5 1/4 floppy, I'd be amazed if you could find the computer that does read it, and that's what...20 years old at most? Let alone the needed words to read the information off of the stupid thing.

This isn't even accounting if we go back further with tapes and cards and then you are well and truly fucked.

[CaptainChewbacca]

I read a few years back that NASA was having trouble transferring all its magnetic tape data to other storage because they didn't have the equipment, and because they didn't have enough people who knew how to use it.

Gives 'data mining' a whole new meaning. Can't wait until my grandson gets his degree in 'digital archaeology'.

[GrandMasterTerwynn]

I like both history and science fiction, which leads me to ponder what sort of things will be around for future archaeologists to study. How long does stored information last? If you leave a thumb drive in a dry place for a hundred years, will I still be able to get the data out? How about 1,000 years? How long do CD's last?

Basically, I'm just opening this thread to ask for any info those who are more computer-savvy may have about long-term data storage methods.

A hundred years is iffy for the mass-consumption CD or DVD. The cheaply made ones will experience de-lamination and subsequent oxidation of the exposed internal layers before even 50 years. Quality CDs, DVDs, and recordable CDs and DVDs may have a life expectancy of up to 200 years.

Flash drives, on the other hand, have different challenges. Prolonged exposure to heating/cooling cycles will eventually weaken solder joints and circuit boards, as will oxidation of the solder joints, and UV-induced (and other) breakdown of the resin making up the circuit board, and the plastic of the case. However, long before any of that becomes a factor, those little transistors making up the flash memory will have long since leaked out all the electrons they've stored (which is how flash memory works . . . to set a bit of data, you stuff some electrons into special transistor which can retain a charge without power being applied,) meaning that if archaeologists dug up a flash drive 1000 years from now, it will be blank.

Of course, even if you did manage to build an electronic storage device capable of lasting a very long time, the data on them will be inaccessible, on the account of the archaeologists not knowing how to build a USB port, something capable of reading from the file system which encodes the data, or the program which formatted the data and presented it in human-readable format to begin with (the specs for these also stored on perishable electronic media, or on printed materials not guaranteed to be printed on archival-quality paper.) Unlike, interestingly enough; the data presented in Egyptian hieroglyphs, aboriginal petroglyphs, and paleolithic cave paintings.

In short, if humans were Raptured away by the Flying RAR Monster tomorrow, alien archaeologists visiting our planet a couple thousand years from now will probably be able to learn more about ancient Egypt than they will about the state of civilization at the present time.

[Ariphaos]

Who said they would be the same interfaces we use today? If I gave you a 1960s-era magnetic tape reel and told you to get me the data off it, how long would it take you?

Not a very apt comparison. Depends on whether or not you wanted me to make sense of the data before handing it over, since the medium is exposed and the mechanisms are fairly simple compared to modern understanding. The data can be made sense of later after it's been properly recorded. This is also true for say, modern hard drives - recovery companies will, if they have to, make a raw photograph of the disc platter if they can't read it electronically and have no suitable head. Even ten thousand years from now, it will be obvious to archeologists that we stored an insane amount of data onto spinning platters.

There isn't a good 'old' analog to modern flash drives, really. There may never be one, either. Technical specifications are printed on all sorts of mediums and stored all over the place.

A hundred years is iffy for the mass-consumption CD or DVD. The cheaply made ones will experience de-lamination and subsequent oxidation of the exposed internal layers before even 50 years. Quality CDs, DVDs, and recordable CDs and DVDs may have a life expectancy of up to 200 years.

I think an optimistic lifespan for a CD or DVD with no label corroding it is about 50 years.

Obviously, at that point there's going to be more interest in data permanence. It's already concerning people now (duh) and the desire to make sure information is available for millennia is probably not going to diminish.

[Count Chocula]

Ghost Rider wrote:

This has to be one of the dumber generalizations of what computers are. If I gave you a 5 1/4 floppy, I'd be amazed if you could find the computer that does read it, and that's what...20 years old at most? Let alone the needed words to read the information off of the stupid thing.

Call me odd, but I DO still have an old 5-1/4 and 3-1/2 floppy drives, and a ZIP, in the garage, along with good old MSDOS and Win98. If I had to, I could be ready to read any of those in about an hour. I may be an exception however.

I voted for flash based on the assumption that, unlike 5,000 year old civilizations, there will be enough continuity of civilization over the next century or ten that the computer technology and programs that we use today will still be available in the future, at least to researchers.

[Ariphaos]

Ghost Rider wrote:
Call me odd, but I DO still have an old 5-1/4 and 3-1/2 floppy drives, and a ZIP, in the garage, along with good old MSDOS and Win98. If I had to, I could be ready to read any of those in about an hour. I may be an exception however.

Are you sure the 5.25" still works? Over my lifetime I've had dozens of them, but all but one eventually failed and then I mistakenly let my sister borrow the computer with that one...

[Junghalli]

I imagine microfilm stored in a durable vacuum-sealed container would probably be the way to go if you want to store information compactly for really long periods of time. As long as it's held in lightless vacuum at a constant temperature I imagine the microfilm could last a long time, and it has the advantage of being stupidly easy to interface with: you just need a microscope.

[Ghost Rider]

Ghost Rider wrote:

This has to be one of the dumber generalizations of what computers are. If I gave you a 5 1/4 floppy, I'd be amazed if you could find the computer that does read it, and that's what...20 years old at most? Let alone the needed words to read the information off of the stupid thing.

Call me odd, but I DO still have an old 5-1/4 and 3-1/2 floppy drives, and a ZIP, in the garage, along with good old MSDOS and Win98. If I had to, I could be ready to read any of those in about an hour. I may be an exception however.

I voted for flash based on the assumption that, unlike 5,000 year old civilizations, there will be enough continuity of civilization over the next century or ten that the computer technology and programs that we use today will still be available in the future, at least to researchers.

So because you have this particular situation it must be obvious future people, or even a specific subsect will be just as prepared with ancient OS and media reading for no other reason then to hope that delicate floppys can be read.

And again, you are basing assumptions on what again? I mean are you even going to try to back up your assertion?

[Count Chocula]

Ghost, with all due respect, YOU brought up the 5 1/4 inch floppy. I did not. In fact, I have little confidence that the floppy disks I have stored in my garage will last another 10 years, let alone 100. My hypothetical tool kit would include a multiple-probe reader, with perhaps various interface ports (assuming our technology, protocols and hardware designs were not lost over time) with legacy OSs and programs in storage (again assuming our technology, protocols and software were not lost over time).

I can't really back up an assertion based on an unknowable future. My blather is simply speculation.

[Ariphaos]

The fact that we use 8-bit bytes is going to be known for tens of thousands of years if not millions. That's plainly available from the structure of just about every commonly used circuit.

If they can recover an HD or CD/DVD, they're going to know byte boundaries, and be able to pick out parity. They're going to know that we made use of 2, 4, 8 and 16-byte values. They're going to know that we made use of floating point arithmetic, and the parts that were the mantissa, etc. They're going to eventually be able to decode a pretty fair portion of - if not the entirety of - the whole x86 instruction set, possibly up through the i386 set and maybe beyond. I say maybe because eventually the advantage of being able to pick apart triply redundant hardened and ruggedized versions of 386 processors is going to be offset by the billions of 64-bit processors floating around.

Your example is positively horrible. A researcher getting a reliable bit image of some sort ten thousand years from now is not going to get 39 or 40 bytes to work with (you are clearly not using parity or ecc there). They're going to get millions. Raw statistical analysis is going to tell a great deal about the individual bytes, even if compressed. Filesystems themselves are also not compressed and at the very least, CDs and DVDs are not fragmented.

Which is especially relevant since I think the most plausible example of such a scenario is using a bunch of unlabeled cds / dvds and sticking them in an oxygen-free time capsule.

[Solauren]

You know, I find this rather funny.

They had an episode of 'The Batman' that featured this. Mr. Freeze up and disappeared on the Batman, so he left all the binary data to restore his old computer system to working condition, or to replicate the software and datafiles, as raised markings on the titanium walls of the Batcave.

Etching onto a chemically no-reactive metal is pretty much the only way to 100% gaurantee that data will remain both readable and unaltered. And that's only if someone doesn't go and melt it down on you, or it doesn't suffer any physical damage.