https://newatlas.com/good-thinking/indent-data-storage/
Cuneiform, the world's oldest form of writing, involved making indentations in clay tablets. Scientists have now developed a data storage system that's like cuneiform on steroids – and it's capable of storing more data than a typical hard disc drive.
The experimental new technology was created by Abigail Mann and colleagues at Australia's Flinders University.
Instead of a clay tablet, the system utilizes an inexpensive polymer film composed of sulfur and a chemical compound known as dicyclopentadiene. Data is stored on that film in the form of a series of nanoscale indentations. These tiny indents are made (and read) using a fine-tip probe mounted on an atomic force microscope ... not by a reed stylus.
In previous attempts at such "indent-based" data storage systems, the indents served as binary code. The presence of an indent represented a 1, while the absence of an indent represented a 0.
Not only were the polymer substrates that were used in these earlier systems difficult to produce, they also weren't very stable or finely workable. That's where the Flinders polymer comes in.
It's sensitive enough that the depth of each indent can be precisely tweaked. As a result, instead of data being stored via two-state binary code, it can be stored via a three-state ternary code in which the absence of an indent is a 0, a 0.3- to 1.0-nanometer-deep indent is a 1, and a 1.5- to 2.5-nanometer-deep indent is a 2.
This capability boosts the system's data density four-fold over binary coding.
What's more, the indents remain intact and readable until the polymer is heated to 140 ºC (284 ºF) for just 10 seconds, thus erasing it. The film can then be rewritten with new data. In tests performed so far, the material remained functional through four write-read-erase-rewrite cycles.
As an added bonus, the indent-writing process can be performed at room temperature, keeping the system's energy requirements relatively low.
"This research unlocks the potential for using simple, renewable polysulfides in probe-based mechanical data storage, offering a potential lower-energy, higher density and more sustainable alternative to current technologies," says Mann, who is a PhD student in Flinders' College of Science and Engineering.
Journal Reference: Probe-Based Mechanical Data Storage on Polymers Made by Inverse Vulcanization, Abigail K. Mann, Samuel J. Tonkin, Pankaj Sharma, et al., First published: 16 December 2024 https://doi.org/10.1002/advs.202409438
(Score: 2, Interesting) by shrewdsheep on Sunday January 05, @01:09PM (5 children)
I don't quite see that. Going from 2 to 3 states gives a relative bit-increase of log(3)/log(2) ≈ 1.6, i.e. 60% increase in bit density. What am I missing here?
(Score: 4, Interesting) by VLM on Sunday January 05, @03:45PM (2 children)
They'll never use it in practice because it'll ruin clocking, but the fourth state is "blank" no mark at all. You can't tell if you lost signal or lost positioning or maybe its broke so IRL most people give a "zero" an actual modulation.
The article oversimplifies to the point of WAY overcomplicating things to the folks who would actually understand it. If you know how modems and RF and stuff like that work, simply imagine the QAM I+Q signal constellation. Usually EE folks like to draw that as cartesian coordinates. Now if you convert X,Y cartesian coords into radius and angle, and make little swipy lines at that angle and of the specified radius, you just implemented QAM on polymer.
Just like modems and RF stuff, once they get the signal to noise ratio high enough they can encode more points (bits) in the constellation by drawing different line lengths and angles. Just like modems they can also implement FEC and some ECC.
It would be hilariously retro if they took something like an old V.32 telco modem protocol and used something like galvos to "write" the I and Q signal into the film. I think it would be hard to move mechanically fast enough but maybe something like a hard disk rotating...
If none of the above makes any sense, time for a bad photography analogy where if you took an absolutely foul out of focus and blurry picture of a weathervane you could encode 4 directions so that's 2 bits of data, a slightly better picture of an analog clock encodes 12 directions that's close enough to call it 4 bits of data, and if you took a really high res HDR picture of a degrees protractor that's 0 to 359 degrees or about 360 directions or close enough to 8 bits of data. Then record pictures with a film movie camera at 24 frames per second and all three signals operate at 24 baud, the weather vane runs 24*2 = 48 bits/sec, the analog clock movie encodes 24*4=96 bits/sec and the high res protractor movie encodes 24*8 = 192 bits/sec. If you used your movie camera to take pictures of multiple weathervanes because they all fit on screen at once, you have DSL-DMT format or multi-channel bonded DOCSIS cablemodems, in a stretched analogy. Classic digital cable, depending on how your local cable company did it, before switched digital video and IPTV, would be another pretty close analogy.
(Score: 2) by tekk on Sunday January 05, @03:53PM
At least in their encoding 0 isn't modulated, it's a true ternary system.
They did claim that they could accurately encode/decode depths of up to 30nm though, so I wonder why they didn't either introduce more states or at least spread out their 2 "on" states to more robust depths than ~1nm apart.
(Score: 3, Interesting) by driverless on Monday January 06, @06:28AM
They'll never use it in practice because you need an AFM to read and write the values. This is a y'all-watch-this, not anything practical - it's a "what cool thing can we do with an AFM", not a new storage medium.
(Score: 3, Interesting) by tekk on Sunday January 05, @03:51PM (1 child)
Remarkably, the full paper is actually linked: https://onlinelibrary.wiley.com/doi/10.1002/advs.202409438 [wiley.com]
The tl;dr is that they didn't think it through. The origin of this claim is that they took the word "secret" in 8-bit ascii, plus one "start bit". Then they looked at it and went "What if we made our own text encoding in ternary." and it turns out that the area needed for 6 bytes in binary is much more than the area needed for 3 trytes in ternary (they encoded each letter as a group of 3 trits.)
(Score: 3, Interesting) by tekk on Sunday January 05, @03:54PM
Sorry, it was my turn to get the math wrong. It wasn't writing 6 bytes, it was writing 6 bytes + 6 bits because of their always-1 start bit, which they didn't bother including in their ternary example. We can say 7 bytes to be round.
(Score: 3, Interesting) by darkfeline on Monday January 06, @09:48PM
This has almost nothing to do with cuneiform. Cuneiform uses orientation and position, not depth.
This is just yet another attempt at superbinary storage with a new substrate.
Join the SDF Public Access UNIX System today!