5 ms·

> 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

> 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!

amalcon 17y ago
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.
iclelland 17y ago
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.
- 3pt14159 17y ago
  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%.
russell 17y ago
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? ;-)
- icey 17y ago
  The byline to the article is "How to store data for a billion years"
  - russell 17y ago
    Apologies. I missed the byline. The writer may be restrained, but the editors/headline writers need a whack on the sides of their heads.
henryprecheur 17y ago
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.

amalcon 17y ago

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.

iclelland 17y ago

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.

3pt14159 17y ago
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%.

russell 17y ago

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? ;-)

icey 17y ago
The byline to the article is "How to store data for a billion years"
- russell 17y ago
  Apologies. I missed the byline. The writer may be restrained, but the editors/headline writers need a whack on the sides of their heads.

henryprecheur 17y ago

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.