Saturday, November 25, 2006

Scam of Indian student developing technology to store 450 GB of data on a sheet of paper

Update:
Mr. Sainul speaks up story here
Today I saw for the second time story coming up in Digg about this Indian student who developed and demonstrated a technology to store 450 GB of data on a sheet of paper. This story was first reported by Arab News and later other media outlets started quoting them.
This shows how technically illiterate the news reporters are.. The idea of storing something in paper is good, but trusting someone blindly when they say they are showing a movie from a piece of paper is absolutely naive.
For people who still believes such stories, please try to find answers for these fundamental questions.

  • How did he make his paper disk?, though the reporter missed to ask this question directly, it seems he used some sort of printer (laser or inkjet). So if we take the best printer available and try to print a digital photo with such a high resolution so that its size is 450 GB, will you be able to print that with out loosing its resolution?
  • Even if a printer is able print at that high resolution, the paper is made of fibre which has uneven surface, think about your CD or DVD being rough like paper, will the drives read it. When you want to store such a huge amount of data, even micron level of difference do matter a lot.
  • Some people where suggesting by using different colours one can squeeze in more data, but what about error tolerance then? This guy is questioning the fundamental reason why digital / binary technology became popular, its because its either 0 or 1 , so its mostly fool proof, we could have used different voltages and instead of binary, use 0 1 2 3 4, but then it will not be fool proof.
  • This guy is claiming he used circles and triangles to store data, so but even for triangle he has to use 3 co-ordinates , that itself utilized 3 bits, so how is he saving space?
  • Barcode companies did their maximum when they tried to develop 2 D barcodes and the maximum they could get was around 2000 bytes of data!!
  • These types of scams happen regularly in India, a guy claimed he got alien cells from rain, someone developed gasolene from leaves in started selling it in large quantities
update 1
Some people have requested me about how many bits is 450 GB, well lets see what Google says...
3 865 470 566 400 bits!!

So he printed these many dots in an A4 paper? even if printed just 1 in 100th of
3 865 470 566 400 ie 38 654 705 664 bits (ie compressing 100 bits to the space of one bit by using different colors and triangles and circles) with a 15 megapixel camera he will be able to detect only
21 037 500 000 bits!!

A4 paper = 8.5 X 11 inch
Using a 15 megapixel camera he can detect 8.5 x 15000 pixels on the breadth side and 11 x 15000 pixels on the length side so total will be 21 037 500 000

update 2: some people have noticed that the arab news article says 256 GB in a paper, and I was saying 450 GB, actually 450 GB came in the register.co.uk story that broke in digg day before yesterday

update 3: I am an Indian and there are many Indians who are brilliant and hardworking, the graduates from IIT are the best in the world!

update 4: I am going to patent this idea, instead of geometrical shapes, we are going to imprint the whole 26 letters and numbers 0-9 on paper, so I m saying it can hold 500 Terrabyte of data on a stamp size paper :) how is that?

Technorati Tags: , ,

114 comments:

Anonymous said...
This comment has been removed by a blog administrator.
Anonymous said...

check this guys 2002 homepgae: http://www.sainul.abideen.dr.ag/

Anonymous said...

Thank you

Anonymous said...

i cannot believe that anybody would accept that 450gb can be stored on paper

Ryan said...

um, a4 ≠ 8.5x11 inches

Anonymous said...

Well he did not say how big of a sheet of paper now did he? ;-)

Anonymous said...

I am an Indian, trust me, lack of basic elementary education (I mean class X) and big dreams can do a lot of stuff.

Do you recall the scam of an Indian making petrol from plants, that was major bullshit and a lot of people accepted it. Why? the reporters who are typically duped are not Xth pass, of have false degrees. The news carries pride and sells papers, so it is easily published without fact check by the editor, this is a typical CNN/Fox like behavior.

It took many scientists to decry the method before it was accepted as a scam.

Dr Ortho said...

I think he is cheating well meaning americans

Anonymous said...

India = pwned.

Multiple times.

Dr Ortho said...

@ ryan
yes my mistake I was using legal paper size, A4 is 8.3 × 11.7. But it hardly makes a difference because area is more or less the same.. Thanks for pointing that to me

Anonymous said...

"So he printed these many dots in an A4 paper? even if printed just 1 in 100th of 3 865 470 566 400 ie 38 654 705 664 bits (ie compressing 100 bits to the space of one bit by using different colors and triangles and circles) with a 15 megapixel camera he will be able to detect only
21 037 500 000 bits!!"

You are trying to take a picture of the entire document at once... Which is absurd. Just like and media reader these days there is a head which reads in fine detail where it's positioned on the media. Instead of taking a picture of the entire document you are taking a picture of that bit only..... You can read all the bits sufficiently fine it just takes time.

Trying to equate data storage technologies to consumer cameras is retarded. Kind of like you.

Anonymous said...

i love lamp!

Anonymous said...

A *lot* of past inventions had overcome people like you... Remember:
Those who say it cannot be done should not interrupt the person doing it.

Anonymous said...

I find it pretty funny that you are preaching about getting facts straight and not believing anything right as you read it, yet you don't have your facts straight! From the article I read the data stored would be about 256 GB. Unless We read different articles then it would be understandable.

Anonymous said...

He can store so much information, because he leaves out all the important words and grammatical conventions of standard english, like so many other Indians.

It's also funny watching the replies from Indian people trying to uphold national sentiment, keep going - you're giving me a chuckle or two.

Anonymous said...

I am the author of the statement regarding camera resolutions and read heads. I can assure you I am very white and you are very much a racist piece of shit.

Here is a moron trying to debunk this and he uses more erroneous logic than the creator of this scam.

Dr Ortho said...

@ anonymous
I got 450 GB from registrar.co.uk article here
http://www.theregister.co.uk/2006/11/23/rvd_system/

Gary said...

Interesting story... pretty far fetched. People will believe a lot, especially if it is presented in a authoritive manner.

http://humanopinion.org (little self promotion).. hey Sameer some students in India may be interested in our £10000 Education Award... (another small plug)

hmmm waiting for the day someone starts talking about how many GB can be stored in DNA.... has to happen... but do we actually want it to???

Daniel said...

The fact that the details gleaned from the whole story are sparse leads me to believe it's a hoax as well.

The register is notorious for hoaxes and without proof of concept, I don't believe a single word of it.

Anonymous said...

~How did he make his paper disk?: It's not a standard sized paper, instead being more like a blueprint or surveyor's map. And why print out the photo when you can just use a compression technique to display the digital data instead of the human representation?

~Even if a printer is able print at that high resolution, the paper is made of fibre which has uneven surface, think about your CD or DVD being rough like paper, will the drives read it.: I have scratches on the surfaces of my optical discs; it's called tolerance.

~Some people where suggesting by using different colours one can squeeze in more data, but what about error tolerance then?: Primary colors with black and white would lessen any effects of color bleeding causing errors, since no tones would have any other distinct color attributes. Quinary would not only provide a compression technique, but would reduce any anomolies to almost nothing using advanced physical detection devices with preventative measures.

~This guy is claiming he used circles and triangles to store data, so but even for triangle he has to use 3 co-ordinates , that itself utilized 3 bits, so how is he saving space?: Instead of a dot we would have a symbol, further used in modern compression techniques. Take hexadecimal. FF takes up less space than 255. A triangle could represent a series of binary digits. Combine that with triangles that are different colors... you know nothing about computer compression and storage techniques.

~Barcode companies did their maximum when they tried to develop 2 D barcodes and the maximum they could get was around 2000 bytes of data!!: And tghey were made during a time when microcomputers were just being introduced into the commercial market! So their standards dictate a reduced resolution compared to today's capabilities! And the have to fit in small areas affixed to merchandise! So they are much smaller in size with a lower resolution! This results in barcodes having a restricted data storage limit (as all data storage formats do) that is very low! Idiot!

~These types of scams happen regularly in India, a guy claimed he got alien cells from rain, someone developed gasolene from leaves in started selling it in large quantities: First, ethanol. Second, apples /= oranges.

Anonymous said...

~Even if a printer is able print at that high resolution, the paper is made of fibre which has uneven surface, think about your CD or DVD being rough like paper, will the drives read it.: I have scratches on the surfaces of my optical discs; it's called tolerance.

You may very well have scratches on the surface of your optical discs, but that's not where any information is stored, so the tolerance you are talking about is the ability of your reader to get past the scratch to the actual data. If you scratched the actual layer where data is stored, you'd have a useless disc.

Anonymous said...

What ! Imperialist nonsensical ! Of course glorious nation India invented superior storage by genius student and will fantastic defeat it's jalous enemys with help of holy president Kalam ! India for eaver !!!

jfperreaut said...

at 1200x1200 dpi in 8 colors on 8.5x11 you'd get 135 meg before FEC (forward error correction)
so this is bogus if he used a standard printer

but if he used a direct laser and a scanning microscope , then he could get a lot higher.. but still 450gb is pushing it

-shodan

Anonymous said...

smazing use of technology and thought.

www.adgerlinux.com

Alex said...

I'd be able to take you a lot more seriously if you knew the difference between 'lose' and 'loose'.

gerryk said...

It is more efficient to store data as packed binary digits rather than geometric shapes. The shapes will have slack space either within them or around them, whereas with raw binary (or trinary, quaternary etc etc) all the space is used by a piece of data.

Secondly, if we look at the amounts of data involved and the sizes of paper... regardless of whether it's 450GB or 256GB, the differences are miniscule, not even a factor of 2, and disregarding the exact size of an A4 or any other page. For ease of calculation, we'll say it's 400GB, represented as HDD manufacturers would have us believe , as 400,000,000,000 bytes or 3,200,000,000,000 bits. Assuming a square piece of paper (for ease of calculation) 1 metre on a side and using binary encoding, we will have a grid of ~560,000x560,000, or each dot ocuppying approximately 2 microns.

These are data densities approaching that of a DVD contained on a medium which doesn't come anywhere near DVD quality tolerances in manufacturing. Surface roughness of paper far exceeds a couple of microns in size, so how does your theoretical read head know what's a bit and what's a paper fibre? The 'tolerance' mentioned by the two posters above is handled by having redundant blocks. The paper storage medium under discussion has no such redundancy, and on top of that, every bit would need more than one redundant neighbour due to the surface imperfections.

Bascially bullshit...

ALM said...

Let’s roll on back to our intro to computer class in high school. Ahhhh…. sweet sweet intro to compys… Anyhoo: we learned number bases there and how to convert between those number bases. Let’s use http://www.uni-bonn.de/~manfear/numbaseconv.php to perform some quick conversion on the decimal value 4444, which requires four digits to express in base 10.

base 2: 1000101011100 / 13 digits.
base 8: 10534 / 5 digits.
base 16: 115C / 4 digits
base 32: 4AS / 3 digits

Now lets stick with base 32 for a while. The digits used to express base 32 are 0-9, and A-V. Using three digits, the greatest value we can express is VVV, or 32,767. The max base 32 value using four number places is VVVV and equates to 1,048,575 in decimal. This value is 236 times greater than little old 4444. Obviously, as the number base being used increases the value that can be expressed increases quickly, and up to this point the terms used to fill the place holder values has been limited to 0-9 and only a portion of the alphabet. An additional limitation on the expression set is the notion that only a black 4444 is the same as a blue 4444. What if it wasn’t?

What if we used color to multiply our place holder tool-kit, and used only moderate technology in determining the depth of our color set? You can walk into any office supply store and pick up a color scanner that distinguishes thousands of colors, but we will limit ourselves to 256 shades. 256 shades of color times 32 is 8192 variations. In base 8192 we need one digit to express 4444. The greatest two digit value that can be expressed in base 8192 is 67,108,863.

If we introduce a truly wide set of symbols for our place holders (for example, Greek letters, the Wing Dings font set, a standard font in upper and lowercase, and a few dozen simple shapes), and multiply that by the 256 color shades, things get interesting. Our place holder set becomes expansive enough to permit base 256*256 or 65,536 distinct digits.

How many of these digits can we comfortably fit on a single sheet of paper? I can easily place an Excel spreadsheet with a grid 21 columns wide by 52 columns tall on a single sheet, or 1092 individual cells.

What is the greatest value of a base 65,536 number with 1092 digits? I have no idea, but I’m positive it will fit on a single sheet of paper, can be printed with a $50 printer, scanned with a $50 scanner, and be translated into binary by a few lines of computer code. I really don’t see what is so far fetched with the idea of storing large volumes of data on paper.

Landincoldfire said...

Hmmm...if this happens to pan out(which it won't). A roll of Charmin tissue will holds loads of data along with loads of other stuff. If one needs to delete data, just flush it. If you need to encrypt data, wipe with it and store in the open. Who will touch it?

gerryk said...

ALM... a large number doesn't equate to a large amount of data

Anonymous said...

ask that person to publish a research paper in IEEE.

Anonymous said...

ALM, thanks for the explanation. You've shown that it's mathematically possible to store huge quantities of data on a single sheet of paper, but that doesn't mean it's technically feasable (sp).

In the mid 90s, I read about an experiement that used base pairs of DNA to solve a computing problem in milliseconds (sp) that large supercomputers would have taken hours to solve. The DNA produced the answer in seconds, but it took weeks of lab work to extract that answer.

All in all, though, it's a very interesting concept.

Anonymous said...

>>ALM

Hey, Mr. Math wizard, it's 65536^1092 or about 3.94e+5259

Wow. You've cracked it. That's enough data to store all of the internet on one sheet of paper!!!

HAHA! Nice one.

Hey, um, good luck scanning in all those colors. HINT: You might have a bit of trouble trying to resolve the symbols printed in WHITE.

Use your brain next time.

andrew said...

You know, this might not be real here, but it's certainly a possibility. If you take a mediocre resolution printer, and 8 x 10 inch sheet could store, at 1024 x 1024 dpi resolution, ~80 million bits or ~10 million bytes. 10 Megabytes. If you include only 1000 shades of detectable color, that jumps to 10 GB. So not bad. So at 450GB or whatever, he's basically saying he can reliably detect 45,000 colors per pixel. Obviously the resolution will change that number, but this isn't outside the realm of possibility, even for a cheap HP printer/scanner combo, once you include some level of error detection and include on the page some reference colors to aid future detection. I think its a great idea with plenty of potential.

Chris said...

My only concern at this stage is that the journalists haven't gone into more detail to make it believable. It could work - if he used a wide enough variety of geometric shapes and a broad spectrum of colours, each shape could represent a larger amount of information than by using single dots.

Anonymous said...

i love desk!

Mipoti Gusundar said...

Mbud. Bud bud bud bud mbud.

Max said...

The best way to look like an uneducated prick is to misspell the word "lose", as in "be defeated" ... as opposed to "loose", as in "your mother". It's LOSING, not LOOSING.

ALM said...

The potential of this process is incredible, but it is entirely natural to suspend belief and doubt it.

"ALM... a large number doesn't equate to a large amount of data" Large numbers are exactly what are stored on our hard drives and memory cards; they are stored in binary. Changing the number base does nothing to the data. The information is still there, but it must be converted back to binary before it can be processed as a useable computer file or program.

As for storing the enitre internet on a single page of paper? It certainly sounds improbable. But... how much binary equivalent data is floating around in the nucleus of the simplest one-celled organism as DNA? Nature surrounds us with simple and extremely powerful data storage.

The real problems with this technology are the optical character recognition, exact color differentiation between multiple printers and scanners, and the fact that the data must be decompressed before it can be used. Wicked cool stuff though.

frankchn said...

ALM, I think your scheme only translates into 17 kilobits.

65535^1093 = 2.5428...e+5264
2^x = 2.5428...e+5264 (find x)
log2 2^x = log2 2.5428...e+5264
x = 17000+

Anonymous said...

This doesn't make much difference to the story but A4 paper is 210 mm × 297 mm, or 8.3 inches × 11.7 inches (see
http://en.wikipedia.org/wiki/A4_paper_size).

8.5 inches × 11 inches is US Letter size and US Legal size is 8.5 inches × 14 inches.

Adam Bishop said...

so lets see.If he uses 3 shapes, and 5 colours, that's 13 possibilities per "dot". (Circle, Square, Triangle, Red, Blue, Green, Black, White). All he has to do, s find another 3 permutations, and he has 16 options per pixel. (for those of you out there that didn't notice, white can't have a shape, as the paper is white.)

If the data to be stored were to be translated into hexidecimal, you can store 1 hexidecimal digit per "dot".

1 Hex digit is equivalent to a nibble, so for every 2 dots you have encoded one byte.

256 gigabytes = 274 877 906 944 bytes. Now, most printers can easily do 1200 dpi. This is linear DPI though, so they can actually do 1440000 dots per square inch. Now, if we assume that we would need at least 9 dots to do all three shapes:

.
. .

. .

. .

.
. .
.

As shown above, that reduces the shape density to 360000 shapes per square inch, or 180000 bits per square inch.

A4 paper (almost foolscap) has an area of 96.6763 square inches, so we can store, using my methods, 17 401 734 bytes, so 16 megabytes, much higher than people here so far have been claiming, and using very, very conservative colour choices and resolutions.

While this is an order of magnitude away from the stated values, this could easily be much higher.

I have assumed a very low resolution (laser printers can easily get up to 2000 DPI these days), no compression, and a very restricted subset of values. I should think it would be easily possible to use 8 bit colour, with no risk of data loss.

Add to this 8 to 14 conversion, or parity values, to ensure data integrity, and I think that what this guy is claiming is within the realms of possibility.

(FYI, using 8 bit colour (256 possibilities):

(3x256)-3 = 765
765 = almost 3 bytes per dot
using 2000 DPI:
2000^2 = 4000000 / 9 = 444 444 shapes/square inch.

444444 * 3 = 1333332 bytes / square inch.
== 128901604 bytes/sheet of A4
== 128 megabytes per sheet of A4.

So as you can see, it's not a case of "is it possible to fit that much data", it's just a case of howdetailed it has to be; add another shape, and the desity per dot goes up massively,

. .
.
. .

maybe?

Is it possible to fit 256 gigs of data on a sheet of A4 with ink?: Yes.
Is it possible to retrieve it?: Possibly, depends how small you go.

If you want precendent, think how small the pits are on a Blu Ray disk are; if we can retrieve a single bit from something that small, can we can surely retrieve something a bit bigger and a bit more detailed.

Anonymous said...

Edit your blog before you publish it. Otherwise, you end up looking like an amateur that doesn't know the difference between "were" and "where" and does not know how to make two complete sentences out of one run on sentence. Amateurish post, at best. Take some pride in your work.

Anonymous said...

I believe someone said A4 was 11.7x8.3. That's 97.11 square inches. A quick search for printers shows some high-quality printers being able to print 9600x9600 DPI, that's 92160000 dots per square inch. Let's just say that he prints in 5 colors, which a computer should be able to distinguish.

Elementary math:
97.11 (sq. in.) x 92160000 (dots/sq. in.) x 5 bits/dot = 44748288000 bits.

That's 44 gigabit. It's not 450 gigabyte, but it is getting closer.

Now, let's just say that the person used some technology that is *new* rather than using off the shelf parts, maybe he can squeeze some extra data in there, and maybe he uses a few times that many colors, and we start to get a little closer to 450 GB.

Assuming you can get a scanner to do the opposite, you have a non-hoax.

I'm not saying that it is likely, but it certainly isn't impossible.

It seems rather premature to either accept of reject this technological "advance" before a technical explanation is presented.

Anonymous said...

Some of the people here going on about using different shapes and more colours should read up on basic sampling theorem.

Particularly Adam Bishop - 9 dots for each shape? Thats nine bits, you fucking tool, ie 2^9 values instead of 7-8 different shapes that you wanted to create with it.

Theres a reason why modern storage mediums use bits!

Anonymous said...

If you look at the picture, it’s much bigger than 1 sheet. So the 450g, even if spread unto 100 sheets of paper is still quite good.
Paper can be made to withstand time better than DVD’s and is cheap & easy to produce.

Now, does it have merit? It sure does! He’s using geometry, which means data redundancy. If you know two angles of a triangle you can infer the third. If you know the lengths of two sides of a triange and one angle, you can infer the rest of the triangle. Even easier with squares.

Data storage? Let’s go with a triangle printed on paper with current inkets.

In a 10×10 pixel grid, an equalateral triangle of, two sides 10 pixels, one angle is 90, two of 45. You can fit two triangles in there.

How much information?
1. The color
2. Length of side 1
3. Length of side 2
4. Length of side 3
5. Angle 1
6. Angle 2
7. Angle 3
8. Border thickness

Times two, because two fit in a 10×10 pixel square. 16 bits.

Then, the square itself! Four angles, four sides, it’s color. Another 10 bits.

Not counting the wasted space inside our square + two triangles.

26 bits, plus a whole lot of redudancy built in thanks to geometry.

In a 10×10 pixel grid. Now, just how big is that when you print at 1200dpi? How many 10×10 grids in 1200×1200dpi for one square inch?

I’ll save you the math, 14,000 grids x 26 bits, divided by 1024 gives us 365k. For one inch!

Assuming we don’t print around the borders, in 8×10 inches, we can then infer 365k times 80, so that’s 29 megabits.

29 megabits / 8 for ASCII storage, gives us 3.6 Megabytes.

So one sheet of paper could easily hold 2.5x floppy 3.5″ disks of binary data instead of text.

With spare room. With data redundancy. With only 15 minutes of my time to think about it.

Now take a grad student that this is his Thesis and he spent many months perfecting. I’m sure he could quadruple my results.

So assuming instead of 3.6 megs per page, it’s really 10 megs per page, it’s still 100 pages for one gig. I’m curious as to how he gets better results, 10×10 grid on 1200dpi is extremely tiny for paper, and most scanners are limited to 1200dpi.

Now I don’t consider compression, because you would compress into a ZIP file before printing your binary in the first place, to save trees right?

Any one of my database tables that’s over 1gig is size compresses to 100megs at max Zip compression. Hopefully this Indian dude is talking uncompressed True numbers.

Now, a paper backup of my database, pasted on my office walls, would give it a nice Matrix feel to it.

Anonymous said...

The funniest thing about all this is seeing the elitists crawl out of the woodwork to bash each other.

To paraphrase some of the comments here and on other sites discussing this issue:

"God you're such an idiot. Go back to basic computer science class and..."

"It's obviously a scam, but you're still an idiot..."

"Do you know how to do math..."

"A4 paper is actually 210 mm × 297 mm... you should check your facts before posting..."

"Dumbass, it's 256GB not 450GB..."

"It's obvious that you know nothing about storage or encryption of data, so maybe you shouldn't speak about it."

"You should check your grammar before posting. Makes you look stupid. Damn fragmented sentences."

For the most part people aren't discussing the real issues like reporters doing basic fact checking or use of critical thinking or just being skeptical of wild-ass claims. Skepticism can be a very good thing, especially in a field where people will lie to you just to get there name in print.

Instead of rudely bashing someone for inaccuracies in their PERSONAL BLOG (we're not talking about the New York Time here), maybe you could politely point them to accurate resources. A few people did this. However, the vast majority just pointed out what an idiot someone was or how poor their command of English was.

I understand how easy it is to just bash someone and not go through all the burdensome trouble of trying to lift them up, educate them, and make better. Why bother with all that? It's much easier and more fun to just call them an idiot.

Anonymous said...

It's late and I'm not doing the math here, but it's definitely possible do better than 2D barcodes by using colors. It's just a reliability tradeoff. 2D barcodes can be read by crappy camera phones at bad angles. By adding a simple 8 color scheme to the 2D barcode you can turn 1 bit into 3 bits. The space that used to be able to represent two values can now represent 8 different values:

BLACK: 1
WHITE: 0

becomes

COLOR1: 000
COLOR2: 001
COLOR3: 010
COLOR4: 011
COLOR5: 100
COLOR6: 101
COLOR7: 110
COLOR8: 111

Obviously a $50 flatbed scanner can do better than picking up 8 colors reliably.

Lets say we have 16K colors. Things get interesting. I don't know if 90-450GB on A4 paper is reasonable.
Maybe they ment 90-450 MEGABYTES, which is believable. You need parity bits for error correction. Magazines could print machine readable music and movies. I don't understand the stuff with the shapes.

It's a cool idea. The thing is, a standard 2D barcode can be used for encoding a URL. In most places Bandwidth is cheaper than ink. And everything is already equipped with hardware to connect to networks. Existing portable devices can already decode 2D barcodes. 2D barcodes are extemely tolerant to poor lighting and different angles due to their low capacity. (just like megnetic cards such as credit cards can still be read even if you don't swipe them as smoothly as a machine can) All the digital cameras I've ever owned, including (maybe especially) those on cell phones compress the images. This causes a serious problem for using such devices to read high density paper media.

Many color laser printers already hide their own serial numbers in every printed page by outputting tiny patterns of yellow dots. This is done to deter bank note forgery.
You can hide data in the low order bits of graphics. You can watermark music by encoding small amounts of data in the low order bits of a highly compressed audio file and recover that data even when it has been played through speakers and recorded off a microphone then recompressed. This is the first time I've heard of someone thinking about encoding large amounts of data as visual images on paper. Interesting stuff.

I don't know if this is commercially viable. But don't call the idea total BS.

algol_i said...
This comment has been removed by the author.
algol_i said...

Besides using color for storing information, you can also use the geometrical figures... and with those figures, you can be a smart guy and draw then into some kind of order that could operate in a similar way as the layers in most modern optical disks technologies. Hum, for example: you can always inscribe the figures one on each order (the most external being the paper border itself)... you can also vary the angles in which some inscribed figures (like the triangle) are related to a preferential reference system (that could be set fixed in relation to the paper boundaries)... imho, it is possible...

mantenham suas mentes abertas... always.

Anonymous said...

Hey, back in the old days we stored information by punching holes in paper - why would you say this is fake?

Anonymous said...

It is not a scam. It can be done. We don't know if 450GB is before the data is compressed or not. Let's assume that it is really 200GB after compression and that's what's being stored. Then let's say we use a 2400x2400 dpi color printer. Assuming 1 bit per dot, that's approx. 720,000 bytes per sq. in. A4 paper is 11.7x8.3 = approx. 97 sq. in. That's about 70MB per page B+W.

Robert said...

Come on you lot, this IS a scam.
You want to see the real thing ...
http://www.adams1.com/pub/russadam/stack.html

Anonymous said...

I think that I figured out what he probably did.

If one 'cell' is 3*3 pixels, there are 512 possible shapes using a 3*3 grid
512 shapes * 256 colors = 131072 bits of data per cell

1200 DPI printout = 400 * 400 cells per inch = 160000 cells per inch

160000 * 131072 = 20971520000 bits of data per inch (2,621,440,000 bytes per inch)

8.5 * 11 = 93.5 square inches

93.5 sq inches * 20971520000 bits per sq inch / 8 = 245,104,640,000 bytes

Therefore, with 1200 dpi printing, 245GB can be stored.

That is, of course, with no compression. A high-quality printer paper would need to be used - ie photo paper, but it seems very feasible.

Eric

Bart said...

His invention is probably not very practical, but your wannebe-computer-scientist math-like conclusions doesn't make any sense either. Bits are zeros and ones, but this goes much further. Algo makes a good point on this too. Symbols, angles and colors (and what else you can put on paper) make good constructors for a data scheme.

Anonymous said...

A lot of people here are making the basic mistake that # of colors = # of bits. # of bits = log2 # of values (colors).

256 colors = 8 bits, not 256 bits.

16.7 million colors = 24 bits, not 16.7 million bits.

Even assuming you could perfectly print and scan 48 bit color, the best available outside of maybe a research lab, that's still only 48 bits per dot.

As for the use of circles and triangles and such, that does reduce data density, but there could be a perfectly good reason for that. Distinguishing individual dots in a print out is dicey at best, because they're scattered somewhat randomly, half-toning, etc. Remember, you only have 4-8 actual colors (inks), the rest are produced by mixing dots. The actual number of pixels you get for a given DPI is generally taken in the print world to be about a factor of 2-3 less.

So, what's the requirement to store 256 GB of data, even before error correction? 256 GB = 2048 billion bits, assuming standard storage convention of base 10 GB. An A4 sheet of paper has 8.3 * 11.7 = 97.11 square inches, so we need to cram about 21 billion bits into a single square inch (edge to edge printing).

If we assume we can actually get 48 bits from color, we still need to store 438 million binary features (dots, shapes, whatever) per square inch, or about 21000 features per linear inch. If each feature is not binary, we might be able to reduce that by a factor of 4 or so, but we'll probably need a hundred times as many dots (for a 10x10 dot feature size) to encode these features reliable.

So I'd say 256 GB is pretty much a scam. Also note that the image in the picture is of a scan, not the actual paper; note the Windows task bar at the bottom. It's a monitor. There's a clearer picture in the original article.

That's not to say this sort of technology wouldn't have its merits, but at these storage densities, it's pretty ridiculous.

For long term storage, you also need to take into account things like color dyes changing colors, etc. You'd be hard-pressed to even get 12 bits of color accurately.

The idea of using a 450 GB picture printed out to a page is actually a good BS-detection technique, as long as the picture is uncompressed. (Don't use JPEG or PNG.) A printer represents information in the same way a picture does, so it's exactly equivalent if you can get a 450 GB (uncompressed) picture printed perfectly accurately on a printed page, and then recover it.

The idea of using a 15 megapixel camera to take a picture of a sheet of paper is pretty stupid, though. The author of this post wasn't thinking clearly when he brought that up.

See the pigeon hole principle if you're going to try and invoke magical compression ratios to solve any problems. Besides, any magical compression is also applicable to other forms of digital storage.

Anonymous said...

The original article (before it hit the Arab News or The Register) can be found here from September:

http://www.deccanherald.com/deccanherald/Sep62006/cyberspace163748200695.asp

I actually contacted the author too, this is not fake. While it might not be possible for 450gb (blame some overzealous journalists for that), he did demonstrate a 45 second video clip based on his algorithm.

That seem's pretty feasible to me and you neglected to mention it anywhere in your blog.

Anonymous said...

As an addendum, multiple signaling voltages are in fact commonly used in high-speed serial buses, memory interfaces, etc.

They don't make sense at the level of individual transistors, where the circuitry to detect multiple signal voltages would outweigh the benefit, but they do allow you to pack much more data into a single transmission line without having to resort to higher frequencies, which is an advantage on external interconnects.

Anonymous said...

Great artcile, BUT:

"update 3: I am an Indian and there are many Indians who are brilliant and hardworking, the graduates from IIT are the best in the world!"

Agreed, all of the Indians I have worked with were very very bright and hardworking and didn't whine like my other american counterparts. They are also by definition the cream of the crop if they make it to the U.S.

As far as schools go however, IIT can't quite hold a candle to Carnegie Mellon, MIT, Chicago, Caltech, Berkeley, U. of Chicago, U. of Illinois at Urbana Champagne... etc.

I've worked with many people from IIT. American schools teach creative thinking, IIT teaches wrote memorization. They make for great engineers to implement, but horrible when it comes to creative problem solving. Sorry, but it is just the nature of the beast.

Anonymous said...

damned jew ass indians

Anonymous said...

Continued...

70MB BW is about 280 MB using 4 colors of CMYK. Now if he can get 256 shades of CMYK, that's 71,680 MB or about 70GB on 1 sheet of paper.

HOWEVER, there are much higher resolution printers and he can use more colors. So it's possible.

Aleksi said...

Yeah but this article is SO WRONG.

15Megapixels != 15 000 by FAR.
15 000 would be KILO.

15Megapixels stands for the total resolution not 15Megapixels wide, 15megapixels tall.

So 15 Megapixels = 15 * 1024 * 1024 = 15728640 pixels, or bits in this case: 1,875Mb of data. But of course, it's not 0 or 1 on a image file :) But it has RGB value for each pixel. Simply add 4 different colors and you already got: 7,5Mb.

You forgot the fact that he's using a scanner not a digital camera, and there accuracy is measured completely different: Dots per inch.

and on printing, even my really old laser printer is able to print 600x600 dots per square inch...
That's 360000 dots total for that area...

Anonymous said...

Well, it does Not have to be a scam - just a better encoding system. It doesn't even need paper - plastic sheet will do nicely.
The combination of colors and shapes can store much more than a binary system of 1/0.

Working like an encryption system, the shape/color encoding system would use a large 'alphabet' of combination symbols, an index, and translation table to get the original 'bits' back out of the color shapes.

It should replace DVDs in about 3 years - because printing a color sheet of plastic that can be read by a scanner is much cheaper than the polycarbonate DVDs and electronic spinning motors of a DVD player.

Cheap printing vs manufacturing DVD/CDs &
A scanner head. Given high enough quality, a fixed scanner head could read a Rainbow Format card like a dollar bill slot on a vending machine.

The Researcher is on to something and don't be surprised if the next story is about him working with Seagate or Hitachi.

Sree said...

Scam as defined here indicates wrongdoing. I dont think this is a scam because, it looks like he has not fooled people and made money in this process. He will not be able to do that also because what he claims is measurable.

My view on Indian scams and Science here.

1TB said...

Why the fuck don't we just store and read data from and in to a BMP or PNG file?
Why the fuck we need the scanner and printer for?

Anonymous said...


This silliness is what happens when we outsource scams.

Anonymous said...

If someone is interested in technology that actually CAN store that much information on such a small carrier, check for instance

http://www.inphase-technologies.com/

who are developing holographic storage. I.e. they use the *volume* of the medium to store more per area unit than any surface-only type normal optical medium will ever be able to.

Anonymous said...

"Why the fuck don't we just store and read data from and in to a BMP or PNG file?"

Maybe becuse there are data already? ;) lol

Anonymous said...

Finally, someone who speaks some sense.

Chui Tey said...

Why not just buy gigs of online storage with a greenback? That's converting a piece of paper into storage media too.

Anonymous said...

Ask anyone versed in theory of information. Using circles and triangle vs dot or no dots will no yield a denser information packing.

Using color may help, but, as you stated it, you have to define tolerance, and define standard for both printing and reasing (so all device have the same calibration)

Also, compare the size of the dots on a DVD (they are truly tiny) and the usable surface, vs the area of paper of a standard page. The 450 GB figure evaporates immediatly.

Also consider that CDs and DVDs use a significant portion of their storage capacity for error correcting, reducing the usable storage space. This is something that would be wise to do on paper.

Finally, I leave the benefit of the doubt for the student named in the articles, because journalists like bold attention grabbing headlines (and tend to be tech unsavvy) it is possible that his claims have been distorded beyond recognition.

Anonymous said...

You ignorant Choot!
There is something called 'encoding'. In my code, if I say 110='Mel Gibson', then I can store Mel Gibson in 3 bits. Your basic fault is that you are strapped down to thinking in bits and bytes and nothing beyond. For your information the Hexadecimal system is much more advanced and efficient than the binary system. So stop trying to push this as a scam and accept the fact that your primitive brain cannot even fathom what encoding he must have used to store something in a triangle. 3 co-ords positioned in a 2-D space can present a theoretically infinite number of possibilities. The limit on this guy's idea is only because of things like paper quality and scanner resolution. In theory, its absolutely perfect.

Anonymous said...

The blogger's complaints may or may not be correct. However they remind me about something else that is true about India: that one Indian will do his best to see that another Indian doesn't succeed, i.e., the frog in the well mentality.

Zugi said...

My guess is that the guy did a totally optimistic best-case calculation of what might be ultimately achievable on paper, advertised those totally unrealistic numbers, and then started developing some software to actually do it and got much, much lower numbers. After all, if he could really do 250GB, he should be able to print and play about 30 full-length DVD quality movies, not one short video clip.

To calculate the absolute best-case data storage on a sheet of paper, consider the best scanners I've seen which can scan 19200 dots per inch with 48-bit (6-byte) color depth. Multiply that by the paper size and you'd get 19200 x 19200 x 8.3 x 11.7 x 6 bytes = 200GB. Maybe he figured he'd do it double-sided and got 400GB.

Now most printers have more like 600 x 600 dpi resolution, and there's no way you can scan numerically accurately with even 24-bits of resolution, let alone 48-bits, and you can rarely print or scan all the way to the edge of a page, so with some more realistic assumptions you can probably hope to store 600 x 600 x 8 x 11.5 x 1 byte = 31MB on a sheet of paper.

My guess is that 10-20MB is the best he's actually be able to do by printing and scanning on even reasonably good consumer quality printers and scanners, which would explain why he's demoing short video clips rather than 30 full-length movies.

If he's using shapes for encoding, my guess is that it's for redundancy and better scanability and therefore actually reduces the amount of data he can store, it doesn't increase it. Tons of mathematical, image processing, and information theory research and billions of dollars have gone into video compression and encoding schemes like MPEG, MPEG-4, DIVX, this guy is not going to come up with something better that uses circles and stars...

Anonymous said...

What's the big deal if the bloke is Indian ? Haven't we heard of American and British con-men either ?
Don't you guys remember Martin Fleishmann and Stanley Pons of the cold-fusion notoriety ?
A cheat is a cheat regardless of his nationality.

Alex Baez [from Chile] said...

ALM I thought it was a different numeral system by just seeing the picture :D But I'm really surprised by how enlightening your first comment was.
Thanks and good luck

Sandeep Z George said...

like i would like to bring to your notice about baud rates and symbol rates. if he's capable of introducing symbols per square inch.. its possible... but the numbers given here are unbelievable

Anonymous said...

Why do you highlight him bieng an Indian? You mention in your update that you're an Indian.

Can't you highlight his scam simply?
Or at some level, are you too ashamed of who you are yourself?

Alan Manuel said...

On compression:
Someone mentioned the pigeonhole principle. It completely debunks the use of innovative compression to achieve those sizes. To those who can't grok it, here's what it means:

Let's say we have a file 450MB in length. Then let's say we have a compression algorithm capable of compressing it by 100, regardless of the input data. Out comes a file 4.5MB in length.

Now, let's compress it using the same algorithm again. Out comes a file 45Kb in length.

Let's do it again. In goes a file 45Kb in length, out comes a file 450 bytes in length.

Now you've compressed a 450Mb file into 450bytes! Whee!

What?? That's not possible?? Of course it isn't! That's the pigeonhole principle! Non-lossy compression algorithms CANNOT compress all streams. They can compress a MAJORITY of streams, but there will always be a set of bit streams that it will INCREASE in length, not decrease. Hence, invoking compression here won't solve the problem.

Yes, a lot of research has been done to compress video and audio. But you neglect the fact that nearly all of these algorithms are LOSSY algorithms - they lose data. That's okay for video and audio - the brain is smart enough to interpolate any missing data. It's not so okay for storing stuff like inventories, though.

Compression won't save you, kids. Not even fancy compression schemes like symbol usage. If that's how you got the 450MB mark, a cd can do the same and store maybe 1G (with compression), and a blank CD can be had for less than a quarter of a dollar in most places.

Scanners:
Here's an experiment. Put a piece of paper with a drawing on it and scan it into your 1200dpi scanner and save it into a BMP file. Then take the paper, put it in the same position, and scan it in again and save it into another BMP file. Compare the contents with any file comparison tool (try GNU diff, with diff -b). If you manage to get the exact same data, thank Allah or Jesus or Vishnu or Buddha or Hare Krishna or Charles Darwin, for you are blessed today.

As a further experiment, do the second scan without even touching the scanner after the first scan. I'm not 100% sure it won't be the same, but I'll be willing to bet maybe 5 dollars or so.

On color depth:
Ink prefers to spread out when it comes into contact with a slightly absorbent material, such as paper. Your printer may print 1200dpi or better, but ultimately the limiting factor here is the gooeyness of your ink. Make it not very gooey and the inks spreads out in semi-random directions, possibly confusing nearby dots. Make it very gooey and it won't leave the print head. Checkmate.

Tom said...

ok. lets say you have a 1. that's 1 bit. Then that one can be 2 colors, black and red for example. A black one is different to a red 1. The shape of the 1 tells us one thing, it's color tells us another. Already that's denser storage.

I'm not a computer science student, and I'm not even sure if 'bit' is the right word to use. But I've not read anything here to convince me that introducing color to the equation doesn't create denser storage. With a degree in maths and/or comp sci, I would expect to be able to build on this very simple example. The jury is still out on this one, and maybe this one guy is a scam, but then again maybe he isn't, but then again, maybe it's inspired someone else to take it further.

And remeber, science is all about disproving things through experimentation. Not simply saying "SCAM" at the top of your voice.

Anonymous said...

Just a quick observation: Im not sure I believe that 450GB could be stored on a small piece of paper, but I dont think this is a scam at all.

What this looks like to me is a novel number system. Binary is base2, because it uses 1 and zero. Imagine a number system of, say, base2048? Where unique geometric shapes are used to represent numbers? Using shape, size and colour as identifying factors, it would be possible to go far beyond even base2048.

Imagine even a 3 digit 'number' in base2048. This could represent 8589934592 in decimal, which is 1000000000000000000000000000000000
in binary. Thats 3 digits to binary's 34.

Just a little something to think about, and the numbers here are simply plucked from my imagination. The number of usable sysbols would be limited only by the printing and scanning resolution of the 'drive'.

Anonymous said...

If you know that a printer with a sufficiently high resolution is required to print the gigabytes worth of data on the paper, then you have a problem.

In the days of dot matrix printers, the printer would buffer and print a line of character data at a time. This could be a stream of normal characters, or a stream of characters mixed with escape codes to force the printer to print graphics. Printing non-western languages was basically an exercise in printing a picture. The top half of the characters of a sentence would be printed as one line, the bottom half on the next line, etc.

These days, the high resolution inkjet and laser printers buffer the entire page. So to print an entire page of data at 256 or 450 GB would require at least that much memory in the printer. Even if such a printer existed, they don't sell it at the local electronic store.

Then you have to wonder how you get all the data over to the printer. Even with fire-wire, 450GB is a lot to transfer. You will probably get better mileage from a tape drive, with an added bonus of a more resilient backup media.

Getting it back in the system would be a challenge also. Back in the days when I worked in the entertainment industry, I set up systems used to scan full frames of film into a computer for CGI editing. I sat and watched the artist that did a lot of special effects for Stargate using this system. The scanning in of the film was done with a large machine that had a cover so that the conditions inside were extremely ideal: correct light temperature, minimal dust exposure, minimal movement of the film they were scanning, etc. The thing is that they had to take at least 3 exposures of the film using Red, Green, and Blue filters (they may have had a Yellow one too, I forget). It took a long time to scan in 12 frames of film at a time. Again, you're looking at a tape backup as a more ideal storage medium for large amounts of data.

This is, at best, a poorly conceived idea.

Anonymous said...

Why would you want to use paper as a storage medium to begin with? So you could fold / bend / tear it and lose all your data? Possible or not, it seems a poor choice that would have limited useful application.

Anonymous said...

Scams are common every where. Dont have to blame India.

Joerg said...

there is/was a German project on optical storage on plain adhesive tape (german ref) reaching a printer resolution of 50000 dpi as a "grey" phase image - i.e. generating a hologram.
Estimating an optimistic byte per dot gives 2.5 GB/square inch. A4 size/inch: 8.3 * 11.7 = 97.11 leads to 243GB per A4 sheet. The storage used a role of tape but it shows that optical storage on sheet substrate could reach this desity.
See also the "milipede" project
http://www.zurich.ibm.com/news/05/millipede.html
reaching 1000 gigabit per square inch, equivalent to storing the content of 25 DVDs on an area the size of a postage stamp.

Brian said...

That he reproduced a 45 second video clip on a piece of paper isn't something particularly novel or new:

With proper video compression, a 45 second video can take up as little as 0.9 MB.

Using .25 inch margins, an 8.5 x 11 piece of paper could store 84 square inches of data. Encoding binary dots at 1024 dpi gets you 10.5 MB already. Drop it to 512 dpi for a little more robustness, and we can still encode our 0.9 MB video clip. Any competent CS student can do that. Now, getting a piece of paper beyond 10.5 MB, that's much harder and requires either much better and robust color scanning or much tinier printing.

Anonymous said...

8.5x11 inches?

Not inches, he said 8.5x11 meters!

jerf said...

Unbelievable how many comments there are here without the "correct" solution.

First, most people have an incorrect understanding of how printers work. Printers do not print "light mauve" or "turquoise"; they print red, blue, yellow, and black, and it's either a dot or not. Other colors are created by dithering the dots really closely together, but still, for any given place on the paper, the color dot is either there or not.

(You can further fake colors by printing the color dots so finely they run together. Good for photos. Bad for data storage.)

So, you've got 4 colors to work with. Assuming a reasonable 300 dpi and total paper coverage of an 8.5x11 (for my convenience), and the complete and total ability to distinguish between every possible page rendering with 100% accuracy, that's (8.5in*11in*300dpi*300dpi*4colorbits) for 33,660,000 bits, or a whopping 4,207,500 bytes total. Multiply by 4 if you go to 600 dpi. Multiply by 4 again to go to 1200 dpi. At this already-absurd level, we've only got 67.32MiB to play with.

You can not exceed this. This is the maximum data that printer can print. The number of distinct printed pages is 2^62,320,000, and that is 62.32MiB. Any fancy tricks you might think you can use will actually *cut down* how many bytes you have. If you try to limit yourself to printing shapes, you grossly cut the amount of legal images you can print. The Pigeonhole Principle, which is anything but fancy math, says that's it.

Scam. Scam. Scam. End of discussion. We're not talking the kind of technological limitation that might be overcome with sheer grit and determination over the objections of small-minded naysayers. We're talking sticking six things in five holes without any of the holes having more than one thing in them. If you still think that's something that you can overcome with sheer grit and moxie, I invite you to spend some time trying it out. Should you solve the problem I can assure you of much fame.

Bungholio said...

I managed to put War and Peace on the head of a pin, however I used my high-resolution Brownie Hawkeye camera to do it. Nothing else would work.

LagunaDave said...

Those who are suggesting using higher number bases are missing the point.

Suppose you use base 256. 1 digit in base 256 stores 8 bits of data.

But you must represent 256 different symbols, and each of them must, obviously, be different from the other 255. What is the minimum number of dots we need to produce 256 unique combinations of them (symbols)? It's 8, exactly the number of bits we are talking about storing.

Whether you store these 8 dots in a line (as a string of bits/pixels), or in some kind of 2-d geometric cell (triangle, square, etc) it makes no difference. If you don't have 8 binary pixels available, you can't store an arbitrary base 256 number.

Now, if I have 256 distinguishable colors, I can store my base 256 number in one pixel, but that's still 8 bits: (2^1) * 8. If I have 2 pixels in 8-bit color, Then I can store (2^2) * 8 = 32 bits.

Let D be the linear density of pixels (pixels per unit length, eg dots per inch) which can be deterministically read and written. Let C be the number of bits of color you can deterministically distinguish pixel-by-pixel.

Then the density of information achievable on a page (bits per unit area)

I = D^2 * C

The first thing to notice is that the color buys you relatively little compared to print/scan density, since the information density scales with the logarithm of the number of colors.

To store 256 GB/page means a target information density (in bits per square inch) of:

J = 256e9 * 8 / (8.5 * 11)

J = 2.2e10 bits/inch^2.

Let's see how close we come with current bleeding edge technology, assuming it would work deterministically, which it would almost certainly not.

D0 = 19200 dots/inch
C0 = 48 bit color

I0 = 1.76e10

Less than a factor of two, but the assumption that you can reproducibly scan 48 bit color at 19200 dpi is pretty far fetched.

As others mentioned, 19200 dpi means a pitch of 1.3 micron - just over 1/1000 of a millimeter. And based on what's in that 1 micron by 1 micron square, you have to distinguish between 2.8e14 possible different colors. But the spacing between atoms is something like 0.3 nm. That means the surface of your 1 micron by 1 micron square contains only about 13 million atoms, and you have to distinguish deterministically between 280 trillion colors. Now you now longer have pixelization (since we are talking about a single pixel), so you do not have the equivalent of 13 million bits of information. Instead, you have to count the number of red, green and blue molecules in your 1 micron pixel to within better than 1/32767 (16 bits), which is statistically impossible when you only have 13 million possible atoms to measure. In addition, if you handle the paper and rearrange ~100 atoms, you are hosed.

So lets back off a bit. What if you can print and scan deterministically at a resolution of 1024 dpi, with a color depth of 9 bits (say 3 bits each of RGB). Very unlikely today, but in 10 years, it isn't unthinkable. Then we have a pitch of about 25 microns, and atomic physics isn't a non-existence proof. Then we get a density of 9.4e6 bits/inch^2 and can fit about 100 MByte in 8.5x11 inches. Unfortunately, though, this is about a factor of 2000 less than was claimed, and not so interesting - which would you rather feed into your computer - 40 sheets of paper, or 1 DVD?

The Average White Guy said...

There's something you're missing here. (except the comment from alm)

You're looking at this project through the bitwise lenses. In traditional computing, a bit is either a 1 or a 0. In this dude's storage system, you expand the possible storage type from 0/1 to shapes of color. If we have 4 shapes, and 4 colors, in any one singe storage unit (where previously we had two possible values) you can store up to 16 different values. I can assume he could be using more colors and shapes than just four of each, so it's not difficult to see how this is possible.

This is an interesting topic. the binary system of memory storage worked because in the "good old days" , a switch was on, or it was off (1 or 0). On a punchcard, it was either a hole or it wasn't (1 or 0) Then when transistrors were gaining speed, it was either an electrical pulse, or it was not an electrical pulse (1 or 0). When magnetic storage (disks/tape) took over it was either magnetically charged or it wasn't (1 or 0). CD's/DVD's? Yep, same thing.

But today, in the world of hi-tech optics and lasers, we should be looking at improving storage, not by reducing the gap between storage units, or by making the storage units smaller (both contribute to higher density - which require finer lasers to read) but by expanding the types of data which can be stored.

LagunaDave said...

To make the atomic argument a bit differently, assuming again that the spacing between atoms is 0.3nm, if you could treat each individual atom as a bit, there are:

3e-7 mm per atom, so
3e6 atoms per mm, or
7.5e7 atoms per inch, and
5.8e15 atoms per square inch, so
5.4e17 atoms (and potential bits) on the surface of the page

To store 250 GByte means that each bit corresponds to creating/measuring the presence or absence of about 27,000 atoms, which is a pretty small amount of ink...

As someone else said, you don't need a printer, you need a scanning tunneling microscope...

Good luck!

LagunaDave said...

Average White Guy:

No, you are missing the point (literally). A shape is an arrangement of dots. Read my discussion of base 256. To store 4 different "shapes", you need 2 pixels. Your shapes are then:

__ , _* , *_, **

Any geometrical shape will require more physical space than the corresponding binary representation in terms of dots.

4 colors gains you a factor of 2 in the number of bits. There is relatively little to gain by increasing the number of colors - every time you double the number of colors, you gain only one bit per pixel.

Can you store data using colored shapes? Of course. But nowhere near the interesting density claimed.

dgi said...

1200 dpi * 1200 dpi * 8.3 inch * 11.7 inch * 8 bits (256 colors which btw. is way too much to be decodable, beacause printers use dithering to achieve larger number of colors) = 139838400 bytes = 133,36 MB
This is a SCAM, sorry.

Anonymous said...

Forget color for one second.

Very simply: You cannot use any encoding scheme that would increase the storage area over binary

Any symbol (triangle / square / absurdly based number scheme) would take up more space to encode the same thing, than you would need in binary of the same area.

If I had a 3x3 square like this:
...
...
...
You have 2^9 possible numbers in binary. If you try to use a special symbol set in the same space they can only take up those dots. So you could have a few symbols in that space, but only if you used every single combination of those 9 pixels could you reach 2^9, and in which case you haven't done anything special, you've just created your own "stupid" binary version.

Now anything that can be applied in terms of color to the "special" symbol set can also be applied to binary.

So even if your "majic" color printer and scanner can do 1200x1200 in 24bit color. The most you can store on a letter size piece of paper is about 385MB

Anonymous said...

A bit difficult to do outside of a controlled environment. Ink will fade, and with that goes your integrity. Good idea, and not denying the fact that it could be possible, but just not the kind of idea for the guy who relies on backed up data as insurance. Lossless not an option.

Have you seen how many horses and bass an Indian can force into a car in south beach Durban?

Eish.

Anonymous said...

It's because all Indians are stupid. I had to work with some bastards from India, oh my god, how stupid they were !!!

Anonymous said...

Stupid is as stupid does my friend.

Pra said...

Why do you post your racist tripe as anonymous, you coward!

So you worked with Stupid Indians eh?

Did you stop to think what they were saying about you when they spoke their own lingo.

Probably something like : 'This stupid asshole couldn't do this, so they got me to do it, and him to take the credit. Plus, he can't even speak the Queen's English. Ha ha."

Yeah, and to the other racist anonymous from Durban, Did you think that the Indians pack the cars with Bass. Huh?

No ways. They just open the door and the fish jump in. So there! Asshole.

Anonymous said...

HI, i think you have one big mistake, in your text. You are thinking about 450GB of dots on the paper. But I think, that if you have dots, you can make many combinations of triangles etc.
I don´t say, that I believe to this message, but you should think about this one small mistake. :-)

Fabian N. said...

I can imagine it works at least in theory:
Each color represents a 16bit hexadecimal value.
The edges of a square-area are responsible for bit Shifting and bool'sche operators within this value, depending on angle and length of the flanks and the adjustment of the object.

The triangles are additionally to the contained color information for the flow control responsible, each square can therefore be "read" in 4 different ways.

The circles determine with their diameter and circumference which codepieces have to be duplicated in certain places, but also mark the beginning and the end of a program.

Anonymous said...

No, it's true. I was hired to work in UK to fix some software that was screwed up by indians. :-) Everyone in that company were laughing about them. Honestly. One guy even told me that he asked one of them "Rakjumar, are you stupid" and he was smiling and didn't say anything. He said that all indians are like that guy. We all laughed about that all the time! :-)

Anonymous said...

Salut,

Si vous avez une info importante à ce sujet, merci d'en faire part ici :
discussion sur l'enregistrement sur papier

Anonymous said...

you know why I hate indians ? Because they don't stay in their own country, but they come to Europe and US and take jobs from white people.

Anonymous said...

450Gb on paper huh? Double sided, laminated, with data stored on both, 1.21 gigawatts of electicity and Mr Fusion should do the trick. I'm still trying to figure out how they made that girl in Weird Science.

Anonymous said...

hi guys

now you can contact this student and ask him about this !!

Sainul Abideen can be contacted at: 0091-98950-81493, Res: 0091-494-2495493,

email: mysainul@yahoo.com

this is the newspaper who reported

http://www.arabnews.com/?page=4§ion=0&article=88962&d=18&m=11&y=2006

Anonymous said...

one thing that comes to mind that i don't see mentioned here is the possibility of a cpu using length, positioning and rotation to store data in addition to shape and color. for example, every shape has a single green dot at one point on it, its relationship to the shape as a whole, as well as its distance to other objects within the paper could all store values, not to mention that the length of each side, the degree of angles and the arc of circles are also data... its all about interpreting it correctly... and that doesn't even factor into the possibilities of laying multiple shapes over each other... personally it is feasible, while it might be impractical.

Anonymous said...

Hey the guy that hates Indians for taking white peoples jobs:

Why do you feel threatened. I live in a country that's black in character -- until you enter the corriders of commerce -- where it is overrun by Europeansm but I don't feel theatended - I welcome it.

If you can't do your job, you boss will find someone else do it.

So wake up and get yourself educated, so that you can keep that precious job.

(And just so you know, US ain't "white people's" country).

Anonymous said...

Respected Sir
Rainbow storage is my work.
Thank you for your kind interest on me and my works.
Please read article on my web page www.kerlontech.com/RandD.html
Pray for me and keep in touch
Sainu

Anonymous said...

Respected Sir
Rainbow storage is my work
Thank you for your kind interest on me and my works
Please read article on my web page www.kerlontech.com/RandD.html
Pray for me and keep in touch
Sainu

Junai Dawson said...

there is someone trying to nick the idea from this gentleman MR. Sainul Abideen in India who is not ready to give out the idea which developed countries can take the advatage and give out in public that they developed by themselves NOW THEY CALL THIS AS A SCAM DONT EVEN THINKING SOMEONE HAS TAKEN AN EFFORT TO BRING THIS OUT...do you yourself believe that you can fly a flight without a help of an India? silly poor ugly creatures trying to pull out from him to steal the idea...thieves...burglers...thieves...burglers...thieves...burglers...thieves...burglers...thieves...burglers...thieves...burglers...thieves...burglers...

message to administrator "IF YOU BELIEVE YOURSELF AS A FOOL AMONG THEM....THEN DELETE THIS"

Junai Dawson said...

I checked Arab News and IT NEVER SAID ANYTHING ABOUT STORING 450 GIGABYTES ON AN A4 SHEET PAPER. All it said is below

"Currently, of the several options available for data storage, DVDs are the best mode. But a high quality DVD, which is very expensive can store only about 4.7 gigabyte (GB) of data. In contrast, the Rainbow Versatile Disc (RVD) can store 90 to 450 GB"

For your kind information RVD is not an A4 sheet paper.

http://www.arabnews.com/?page=4§ion=0&article=88962&d=18&m=11&y=2006

Where did the blogger find these details from arabnews?

If have eyes read this and apologize the readers...

Anonymous said...

I am one of those who have actually visited Sainul Abideen to know the truth behind the allegations against him. HOw many of you have done that, for a change? It seems that he has never raised a "450 GB on an A4 sheet claim", and that was some thing that I have verified. Besides, all of his claims have been misinterpretted. I was fortunate to see his product in person, in working condition. Besides, he has shown me an apology letter sent by Mr Chris Mellor (Author of the article that started all the rumours against him) for spreading rumours on his blog. The Hypocrisy of this author is to the extent that he has never pulled back the article from the blogs, nor has he publicly apologised. Shame on the way these so called 'critics' have taken sides against the new developments in third world countries. And why is such a stigma attached to the innovations done by students of third world countries?

Diana said...

A hoax is an attempt to trick an audience into believing that something false is real. There is often some material object (e.g., snake oil) involved which is actually a forgery; however, it is possible to perpetrate a hoax by making only true statements using unfamiliar wording or context (see DHMO). Unlike a fraud or con (which is usually aimed at a single victim and are made for illicit financial or material gain), a hoax is often perpetrated as a practical joke, to cause embarrassment, or to provoke social change by making people aware of something. sportsbook, Many hoaxes are motivated by a desire to satirize or educate by exposing the credulity of the public and the media or the absurdity of the target. For instance, the hoaxes of James Randi poke fun at believers in the paranormal. The many hoaxes of Alan Abel and Joey Skaggs satirize people's willingness to believe the media. Political hoaxes are sometimes motivated by the desire to ridicule or besmirch opposing politicians or political institutions, often before elections. http://www.enterbet.com

Anonymous said...

Really nice information, thanks for sharing!thanks!
_

And... It only took until the send poster to fall for that one. Yer so smart.


8

IT Soup Chef's top picks