Scientists at Microsoft Research in the United States have demonstrated a system called Silica for writing and reading information in ordinary pieces of glass which can store two million books' worth of data in a thin, palm-sized square. In a paper published today in Nature, the researchers say their tests suggest the data will be readable for more than 10,000 years.
What tiny pulses of light can do:
The new system, called Silica, uses extremely short flashes of laser light to inscribe bits of information into a block of ordinary glass. These pulses are called "ultrashort" for a reason. Each one lasts mere quadrillionths of a second (aka femtoseconds or 10^–15 s). To get your head around that: comparing ten femtoseconds to a single minute is like comparing one minute to the entire age of the universe.
Writing in glass:
Femtosecond laser pulses also have a practical technological application. They can be used to make changes deep inside transparent materials such as glass. These lasers produce light of a wavelength that normally passes through glass without interaction. However, when ultrashort pulses of this light are tightly focused on a particular region, it produces an intense electric field that alters the molecular structure of the glass in the focal zone. This means only a tiny three-dimensional volume, often less than a millionth of a metre to a side, is affected. This is called a "voxel", which can be made at precisely controlled positions in the glass.
[...] The Silica project does not claim to have made a new scientific breakthrough. Instead, the team presents the first comprehensive demonstration of a practical, real-world technology. Their work brings together all the key elements of such a storage platform based on femtosecond lasers and glass. It includes encoding data, writing, reading, decoding and error correction. The work explores different strategies for reliability, writing speed, energy efficiency and data density, and involves systematic assessments of the data lifetime. These allow an extremely high storage density of 1.59 gigabits per cubic millimetre.
[...] Finally, accelerated ageing experiments suggest that the written data, even in the case of the more sensitive phase voxels, could remain stable for more than 10,000 years. This vastly exceeds the lifetime of conventional archival storage media such as magnetic tape or hard drives.
[Journal Reference]: https://opg.optica.org/ol/abstract.cfm?uri=ol-21-24-2023
(Score: 5, Informative) by owl on Tuesday February 24, @05:12PM (16 children)
So you have your fancy glass substrate with carefully created "voxels" inside.
In order to access the data in your bit of glass, you are going to need a "glass drive" (for want of a better name). You insert your little glass slide, the "glass drive" reads the data from the glass, you are good.
But then, just like with almost every other "removable storage system" that has ever been developed, after a short time, someone's going to create the "next great thing". And if that "next great thing" end's up supplanting your "glass drive" system to such an extent that sales of "glass drives" fall off a cliff (i.e., floppy drives, CD drives, DVD drives, etc.) and the makers of "glass drives" start winding down their production lines (because few are being sold), then you end up in the same boat as with all those others. You have a stack of 3.5" floppies, but can't find a 3.5" floppy drive to read them, you have a stack of CD's, but can't find a CD drive to read them, etc.
The glass slide might store your data for 10,000 years, sure.
But if in 12 years you can only find used "glass drives" on eBay, and in 12 years after that you can't find any working "glass drives" anymore (and can no longer obtain one or more critical components to effect repair if you wanted to do so) then a 10,000 year data lifetime isn't very relevant if, 24 years later, you can no longer find a "reader" for your glass slide.
(Score: 0) by Anonymous Coward on Tuesday February 24, @05:42PM
If there is a monetary reward in doing so greater than the cost of recovery.
"The quotes from Chairman Gates" will be available in perpetuity.
(Score: 3, Funny) by krishnoid on Tuesday February 24, @06:50PM
Or they slowly add stuff to the glass as it makes it into the public domain, put the reader on the Internet, and permanent (?) storage and access.
(Score: 4, Funny) by GloomMower on Tuesday February 24, @07:28PM (8 children)
You mean because you don't want to buy a drive. It fails because it didn't come built in to your tablet? You can get 3.25 inch floppy usb drives online for $20.
(Score: 3, Touché) by HiThere on Tuesday February 24, @10:37PM (7 children)
Yeah, but is the data still readable? I've had lots of floppy disks fail.
Javascript is what you use to allow unknown third parties to run software you have no idea about on your computer.
(Score: 5, Insightful) by vux984 on Wednesday February 25, @12:04AM (6 children)
Yes, these would still be readable. That's the whole premise. That is precisely why this drive tech is better than floppy disks - the data will be readable for thousands of years.
The OPs complaint is that tech is moving so fast that the drives to read/write them will be could be hard to find in mere decades, let alone centuries or beyond. This is a pretty valid concern. The drives may cease to be available, the computers they are compatible with may disappear, the technical data and documentation on how they work may be lost. The means to interpret the data itself may be so far removed from today that even if we could read them we might not know what it says.
We have many historical manuscripts today that we can see plainly, and we don't know what they say. The language they are written in is unclear or has been lost. Fast forward a few thousand years from now to the 71st century, and the idea that we're going to be able to read 21st century english is far from assured. The notion that we're going to be able to decode 21st century english stored in a file saved in a Microsoft office format... or recorded to an audio/video codec that's popular right now is even less assured. The notion we're going to be able to pull that data from some 21st century storage device, correctly decode it, and understand it ... is a massive leap of faith.
Clearly a data storage medium that can actually reliably last that long is a prerequisite to pulling that feat off, but there's a LOT of other prerequisites, and I'm not convinced finding a reliable storage medium is even hard compared to the others.
(Score: 2) by Uncle_Al on Wednesday February 25, @12:59AM (1 child)
"a data storage medium that can actually reliably last that long is a prerequisite to pulling that feat off"
and what would be of value would you record?
where you buried all of your nuclear waste?
how many times Trump told a lie?
(Score: 3, Insightful) by vux984 on Wednesday February 25, @01:31AM
I don't think there is question about the potential to put something of value down. Some things will have greater value, others less so, but archaeologists today would be beside themselves with glee if they discovered a few weeks worth of low quality fox news articles and its users plainly ignorant comments about a local 5th century political conflict.
But of more general value, the history of the world as we see it today would be of tremendous value to a future civilization that cares about such things.
A snapshot of the internet, as it's already rotting away; and wikipedia, and our scientific progress; theories, industrial and construction processes, mathematics, and medicine - whether they've advanced far past it or regressed behind it ... could also be of immense interest and value.
... if they can read it.
(Score: 2) by HiThere on Wednesday February 25, @04:17AM (3 children)
Text can be decoded. It's true it can take a large corpus, and even then it's not guaranteed unless you've got some way to decode part of it, but it's been done. There are exceptions, e.g. we don't really know Mayan, and there's a reasonably large corpus. But Hittite has been decoded reasonably well.
If there's a durable medium, intermediate steps can be recorded, and you can use one to decode an earlier one. And something optical can, in principle, be read with a microscope. (I knew someone who did that with mag tape, but he needed to "develop it".)
OTOH, a complexly formatted file will probably just be lost. So XPM might well be preserved, but not png.
Javascript is what you use to allow unknown third parties to run software you have no idea about on your computer.
(Score: 2) by hendrikboom on Wednesday February 25, @03:22PM (2 children)
Has anyone ever managed to figure out a lost language without some kind of Rosetta stone or some kind of similarity to an already understood language?
(Score: 2) by HiThere on Wednesday February 25, @07:35PM (1 child)
You always need SOME entry point. But if there are lots of permanent records, that shouldn't be a problem. The problem is if all the intermediate forms disappear.
OTOH, it's NEVER complete. Not even between different modern languages. Too much depends on cultural context.
Javascript is what you use to allow unknown third parties to run software you have no idea about on your computer.
(Score: 2) by cereal_burpist on Friday February 27, @04:39AM
https://en.wikipedia.org/wiki/Voynich_manuscript [wikipedia.org]
Future
civilizationAI: "Those stupid humans from the late Holocene! Wasted our time, just for some images of domesticated felines."(Score: 4, Funny) by higuita on Tuesday February 24, @09:28PM (2 children)
Well, to archive data, even obsolete tech is used if needed!!
we still use TAPES to do backups, they are "cheap", easy and can last ... some time! due to the lack of better replacement, tapes are still in use and believe me, they are MUCH older than 12 years ( IBM 726 tape drive introduced at 1952). And yes, tapes are slow, hard to manage, but disks are too sensible, ssd expensive,small and rot, CD/DVD/blueray rot and too small... so tapes are still a good solution
So if this works, if it can store data fast enough, store it for a LONG time and be able to retrieve it... people will replace TAPES, of better, add "glass drive" to the backup system and it will keep being used for a long time
(Score: 0) by Anonymous Coward on Tuesday February 24, @11:35PM (1 child)
Tapes require physical contact to the read/write head, which means the heads will wear
out and are expensive to replace.
Google was (is?) chewing through LTO drives because of the sheer volume of their
backups
(Score: 2) by higuita on Monday March 02, @10:24PM
True... but they are "cheap" too and you usually rotate tapes, so you keep data fresh and you know you can reuse a take X amount of time before discard or demote to a less important backup role... some backups are just for compliance, so as long you have, say 2 copies, you are fine, you will probably never read them anyway
(Score: 2) by VLM on Wednesday February 25, @01:14PM
Possibly, they could record in a very non-optimized method.
Nobody owns the microscope equivalent of a read-write head for casual magnetic media inspection and cds and dvds are hyperoptimized.
However, this thing is basically a glass engraver so they could just "print stuff" like printing to a plain old laser printer. Hopefully they'd use plain text and not unreadable QR codes. But in theory they could make it very much like higher density microfiche and just slap the glass block under a decent microscope to read it.
I'm just barely old enough to remember microfilm at the local public library and IIRC I was "mostly sorta" able to read microfilm with a magnifying glass. It would not be comfortable but it was "good enough" to see the title of a newspaper at least to make vaguely certain you got the correct roll and some dumbass didn't wind it on another spool backwards or something.
Given an infinite amount of storage they could engrave both plain text and QR-like codes of UTF-8 text. QR codes can be almost 50% efficient, a roughly 200x200 pixel QR code 'shoud' store 5K but actually only stores about 2K bytes or probably well over a thousand UTF-8 glyphs. A standard IBM punch card stores IIRC precisely 960 bits of data so I was fooling around with the idea of printing QR codes for retrocomputing applications. Annoying density. If you figure each QR code is bigger (physically) than a 2K eprom but stores about as much data as a 2K eprom it would be quite realistic as a stunt to build a machine that plays 1980s home computer and video game cartridges off a paper sheet scanner. In my infinite spare time of course. QR codes on paper doesn't store enough to be "interesting" WRT 80s home computers or 60s industrial computers but it enough to do crazy things, well, in theory.
Back in the day there were infinite attempts at getting enduser consumers to use barcodes, usually unsuccessfully, but for error detection and recovery of OCR text maybe that would be "good enough".
(Score: 3, Insightful) by hendrikboom on Wednesday February 25, @03:51PM
Since technological change may make it difficult to read after a while, it's still prudent to copy the data now and then to new media, just as it is with today's storage media.
This is as true if the medium disintegrates as it is if the machinery to read it disintegrates.
(Score: 1, Insightful) by Anonymous Coward on Tuesday February 24, @05:16PM
Or maybe that disparages the old, pre-Disney muppets too much.
Either way, this is marketing disguised as research. MSFT throwsd a few dollars at some starving researchers and then gets the company name prepended to all subsequent reports.
(Score: 2) by turgid on Tuesday February 24, @05:50PM (17 children)
I thought that glass is a fluid and that's why windows in buildings that are hundreds of years old go wavy because the glass gradually deforms under gravity?
I refuse to engage in a battle of wits with an unarmed opponent [wikipedia.org].
(Score: 5, Informative) by istartedi on Tuesday February 24, @06:26PM (6 children)
It's an urban myth. Old glass is wavy because earlier flat glass making processes were less precise. Glass blowers took boules and spun them to get something reasonably flat. Perhaps they used metal rollers or other tools; but it was artisanal. Any glass made with a modern process won't look like that, even after 100s of years.
Appended to the end of comments you post. Max: 120 chars.
(Score: 5, Informative) by krishnoid on Tuesday February 24, @06:48PM (3 children)
Recently tested [ceramics.org], it seems.
(Score: 2) by turgid on Tuesday February 24, @09:14PM
Well thank you both very much. That's very interesting and informative.
I refuse to engage in a battle of wits with an unarmed opponent [wikipedia.org].
(Score: 5, Interesting) by higuita on Tuesday February 24, @09:33PM
>The team’s calculations show that the medieval glass maximally flows just ~1 nm over the course of one billion years.
well, that confirms that they actually really move, only much slower than expected!!
this is still a problem for this glass cube, but maybe then is why they report 10,000 years and not 100,000 years
IBM did build disks in a glass substract, the "deadstar" disks that suffered from this problem...but they also found that data moves in pack, so later software updates could compensate for that for a much longer time
(Score: 3, Insightful) by VLM on Wednesday February 25, @12:46PM
They went to a lot of work compared to looking thru eyeglasses or a telescope.
You can measure much less than a 1/10th of a wavelength on a telescope mirror and optics pretty much completely stop working if they're messed up more than a quarter wavelength over a large scale (like a sag) and a quarter wave of blue light is pretty small, so the hilarious claims of building windows sagging 1/4 inch in a century mean glass optical products would only work for a very short time.
A tenth of a wavelength of blue light is handwavy 2 millionths of an inch. And a quarter inch of saggy church window glass would be 125000 times more than 2 millionths of an inch. So if it took a church window 1000 years to sag, eyeglasses would be worthless distorted puddles 125000 times faster, or ... what about three days after manufacture? Anyone who owns eyeglasses that have worked longer than three days knows the church window claim is therefore BS.
Another problem is there is no "glass" atom element like there is an "iron" atom element. There's probably some way to make super cruddy "transparent material" that sags like some plastics do in sunlight. But its possible to make in theory or at least as a thought experiment is a LONG way from "all glass sags 1/4 inch per 10 centuries" or whatever total nonsense.
(Score: 3, Interesting) by aafcac on Tuesday February 24, @07:26PM (1 child)
That sort of makes sense given that the "flow" itself isn't always top to bottom. The windows in my parents' house growing up had ridges in them that were aligned top to bottom, rather than parallel to the bottom. Which isn't at all consistent with the flow of glass downwards.
(Score: 3, Informative) by Anonymous Coward on Tuesday February 24, @08:17PM
The myth started because a lot of those old windows were thicker at the bottom than the top. Thing is, old builders weren't idiots and when given an uneven pane of glass to install they put the thick bit at the bottom.
(Score: 3, Interesting) by Rich26189 on Wednesday February 25, @02:28AM (2 children)
This could be tested by putting panes of glass in a centrifuge like device and spin the panes at high speed such that the centripetal (centrifugal?) forces are parallel to the plane of the glass. Though it still might take a while, e.g. the pitch drop.
(Score: 3, Interesting) by Anonymous Coward on Wednesday February 25, @02:48AM
a REALLY bad idea
there was a reason 64X CDROM drives didn't literally survive very long
when the disks experienced rapid disassembly
(Score: 2) by VLM on Wednesday February 25, @12:59PM
You could just use a centrifuge and test tubes and put something with a higher density than glass in the tube and see if buoyancy takes its natural course. It will not, of course.
Glass isn't very dense compared to metal. Liquid mercury or liquid gallium would do.
There's at least one hypertoxic organic chem specialty solvent made with bromine thats denser than glass so if you filled a test tube of flexible glass with it and spun it, the glass would tend to float on top and it would create an impressive environmental nightmare mess. A glass microscope slide will float on it. Seen the demo in a uni classroom. Taking a wild ass guess and some intuition its probably some variety of brominated methane or maybe a longer brominated hydrocarbon. IIRC if you disinfect water using bromine or disinfect water containing bromine you get some of this in the water, like out in the world, so its not an exotic compound but being a disinfection product as you'd guess its not very compatible with life. Drinking halogenated water seems to be a pretty bad idea and centuries from now people will probably look back at it like we look at drinking radium water.
(Score: 0) by Anonymous Coward on Wednesday February 25, @04:04AM (6 children)
No one seems to have mentioned that glass is brittle. The goon with the big wrench who gets your password by threatening to whack you? That same goon can also shatter your glass memory. Or, less dramatically, if you drop your glass memory on a hard floor (ceramic tile?) it's gone.
(Score: 4, Informative) by HiThere on Wednesday February 25, @04:24AM (2 children)
Not all glass is brittle. And even for soda glass, how brittle it was depended on the size of the piece vs. it's thickness. Really small pieces aren't that brittle. Something half the size of your little finger nail and 1/8 inch thick wouldn't be that brittle. (Make it too large, though, and even if it's not brittle it cracks under thermal stress. But consider "glass bricks"...it can be hot inside and below freezing outside, and the glass brick doesn't care.)
Javascript is what you use to allow unknown third parties to run software you have no idea about on your computer.
(Score: 0) by Anonymous Coward on Thursday February 26, @03:58AM (1 child)
Agreed, the shape and size matters. But that goon with the big wrench can probably shatter most sorts of glass.
Slightly related story??
I thought the glass on our electric range top was pretty tough, since it survives local heating by the red hot coils in the individual "burners", while staying at room temp only inches away. Not so, I dropped something heavy on it by accident and it shattered like window glass with lots of sharp pieces. Just because it has a temperature coefficient of expansion very close to zero (over the operating temperature range), doesn't mean that it's also shatterproof.
(Score: 2) by cereal_burpist on Friday February 27, @04:09AM
That made me think of Pyrex [wikipedia.org] cookware. I didn't know there are two different brands (pyrex and PYREX) and two different compositions (soda–lime glass and borosilicate glass).
(Score: 0) by Anonymous Coward on Wednesday February 25, @09:44PM (2 children)
Use laminated glass
(Score: 0) by Anonymous Coward on Thursday February 26, @04:01AM (1 child)
> Use laminated glass
As used for car windshields? Take a scenic side trip sometime to your local auto recycle/junk yard. Lots of windshields held together by the central plastic lamination, but glass on both outside and inside all cracked from a crash.
(Score: 0) by Anonymous Coward on Thursday February 26, @09:19PM
Yeah, but a little bitty square probably won't break when dropped on the floor.