I congratulate you, sir, on being clearly smarter, having more initiative, and generally being an all-around better employee (and human being!) than the obvious waste of skin you're responding to.
Given the successes they are listing, I'd be a little worried. Jose Padilla struck me as sort of a failed terrorist wanna-be, and the plot to destroy the Brooklyn Bridge was a non-starter (Iyman Faris quickly realized that their idea was impractical, and called it off).
I wouldn't be surprised if there are other, similar scale plots being averted that don't make the news. But if something significantly bigger or catchier was disrupted, I think Der Prez would think it was more important to keep public support for his activities than to maintain operational secrecy. So he'd go public in a big way.
There is something to be said for efficiency and information-sharing within the intelligence community. But there was also something to be said for the limits on that sharing. Is the FBI really respecting its suspects' civil rights if it can call up the CIA to get information that they can't legally obtain?
It's a tricky question, and I have no idea how to thread my way through that particular landmine. Obviously, with an organization like al-Qaeda, the line between foreign and domestic investigations gets more than a little blurry. But I don't think it's clear that the race to tear down the wall between the CIA and the FBI has resulted only in "more efficiency."
I never said it wasn't a freakshow over there. But I double-checked a couple of the citations, and the laws are very much still on the books. Which was the whole point. I'm not saying it's right or good, or supporting their website, or anything else. I'm just correcting a factual inaccuracy.
The decision to suspend the student was made by a public school. As in, by people representing the government and providing a government-sponsored service. If a school can punish you for expressing an opinion outside of class and off campus, then you're effectively saying that the government is allowed to punish people with opinions it doesn't like.
Public schools have an obligation to educate every child. While they have control over what a student does while under their care, they cross a huge line (one supported by numerous judicial decisions).
Ahem. I think it would be more correct to say that no state enforces adultery laws. Good thing, given that about half of married people cheat at some point.
I get pretty frustrated with Yahoo Answers myself. It takes the longest time to find questions that are serious, interesting, and relevant to me.
On the other hand, the barrier to entry for Google Answers was way too high. You had to pay to ask, and you had to go through a small job interview to answer. Once I found that out, I never touched it again. It wasn't free, so it never developed a community around it. Google should have seen that coming a mile away.
It seems like the way to go would be a two-tiered system. People would be able to ask and answer questions, and eventually if they generate a high enough "trust metric" they would be allowed to answer for-pay questions. People could ask questions for free, or chip in a few bucks to motivate answers. People with insufficient credibility would be allowed to answer as well, but they'd get the "anonymous coward" treatment (e.g. answers not visible by default). Once the question is closed, the person has to select the best answer(s), and the money is divvied up.
Yahoo should learn from World of Warcraft: You can get people addicted to leveling up.
Of course, once you get money involved, people will start looking for ways to game the system.
The program he describes is one that the FISA courts wouldn't think twice about approving surveillance on. In fact, if we already know that a phone call has a "known terrorist" on one end (someone they're already surveilling) then they should already have the warrant to listen in on the call, regardless of who is on the other end of the line or where the call is being placed from.
No, the only way the program makes sense is if they're randomly listening in on any phonecalls between the U.S. and "evil terrorist countries".
Okay, yes, the last sentence is correct, in its own saggy, strawman way. Not many people are claiming that they're listening in on entirely domestic calls, so where the post you're congratulating is correct it is also irrelevant. The important thing is that you need a warrant to listen in to calls where one party is in the U.S. Which the Bush Administration is not getting. Which makes the whole program illegal from top to bottom.
Keep congratulating yourself for going against the Slashdot groupthink. But be warned: sometimes the groupthink you rail against is nothing more than reality.
If by "terror suspect" you mean "person who lives outside the United States," then I can see how you make a lick of sense.
Otherwise, given that the Justice Department has steadfastly refused to give any details on who is being monitored (to avoid "aiding the terr'ists") you don't know who is being monitored, or for what reasons. You have no way of gauging their decisions on who should and shouldn't be monitored. You have no way of gauging whether anyone's civil rights are being violated. You have no way of gauging whether the people running the program have valid probable cause for being suspicious of the people they're listening in on.
All you have is the promise of your president that the people we're spying on are bad people or are talking to bad people. But if they already know that, then why can't they just get a warrant? I can see only two possible reasons. The first is that these wiretappings really don't pass the sniff test. The second is that the Executive branch no longer feels constrained by the Judicial branch. I leave it to you to decide which is scarier.
There have been 10 midterm elections from 1970 on. This is the third biggest*, and two of those are unarguably historic events (the midterm surrounding Watergate and the Republican Revolution of 1994). I would describe a change of ten or eleven seats as normal.
Looking at the historic data, you see much bigger shifts in the elections before 1974 than in the elections after. Average before: 30. Average after: 18. Which indicates that the incumbency effect has been getting stronger lately. Throw in gerrymandering, and it's obvious that this election was well outside the norm.
* Maybe fourth, if you're going by the % of seats available.
By "engage" he means "start talking to them and pointing out how its in their best interests to stabilize Iraq." I really don't think he means "start dropping bombs."
Unfortunately, it doesn't work. What you're proposing is some sort of lossless compression scheme. If true, it would be way bigger news than some guy applying ink to a fibrous substrate. It would also violate about eighteen basic principles of information theory.
As I've said elsewhere, the whole "shapes" thing is bogus. In the end, all you really have is the dots, and shapes just end up bringing regularity to the patterns of dots. In information theory, regularity is the opposite of entropy, and the entropy of a pattern is identical to its ability to carry information. Quick version: 11111111111111111111111111111111111111111111111... has much less entropy than 10001011110100011101011110100001011010101000011110 0001... and the latter carries much more information (because the latter message cannot be compressed as small as the former message).
In the case of shapes, knowing that your dots form either "circle", "triangle" or "square" means that if you read a given pixel, it is possible to say something about neighboring pixels. In order to convey the most information possible, knowing the contents of pixel A shouldn't tell you anything about any other pixel on the page.
Here's a challenge: Name me the country that has fewer controls on its health care system, yet delivers better health outcomes. I can name plenty of countries where people live longer, and plenty where less is spent per capita. But in all those cases, the government is more in control of how health care is distributed, not less.
Come to think of it, name the free market paradise where the food industry is both lightly regulated and primarily devoted to delivering the sort of food that will keep people healthy, instead of clogging arteries and turning its citizens into porkers.
I have little hope that this free market utopianism, if implemented, could do anything beyond lining the pockets of the rich. Meanwhile, every country I perceive as having a better quality of life has a more stringent approach to regulating business than the U.S.
While it's not easy to compare infant mortality rates between the two countries, it's safe to say that Cuba has an impressive health care system. Cubans live about as long as we do, even though we spend more per person on health care than Cubans spend on everything put together.
Further, unless Cuba is flat-out lying about their infant mortality numbers (not just adding them up differently), they're doing stunningly well with regard to infant mortality. The U.S.--which has far greater resources to tackle the problem--is a pathetic underachiever by comparison.
Some of America's poor performance may be attributable to different ways of collecting statistics, and to more active intervention in the cases of very high-risk births. But you're ignoring the role played by inadequate prenatal care, poor nutrition, poverty, and inadequate access to health care. It's hard to respect such a Panglossian attitude when there are plenty of better outcomes staring us in the face. Cuba, for its many faults, shows what can come of a government that is devoted to ensuring that everyone has access to basic medical care. The United States, for all its wonders, delivers terrible health outcomes at exorbitant rates, because it marries an obsession with technological gadgetry and a complete lack of interest in making sure that everyone has the basics.
Unfortunately, I don't think that works. The only way to achieve higher-order patterns is by bringing some sort of regularity to the lower-order patterns, which decreases the overall entropy (and hence the maximum amount of information that can be conveyed).
That's a completely different issue. Eugenics aside, we have no control over the distribution of the talents you listed. But as a society, we have a lot of control over how wealth is distributed.
If you're trying to say that some of these attributes just naturally lead to making more money, then I'd agree. I'm fine with the fact that one person can earn more than another. What I have a problem with is income that is not earned: income generated entirely by ownership of some business, property, etc. I'm an unrepentant communist, in the sense that I believe that the wealth generated by a business should go to the people doing the work, and that mere ownership confers little value.
All I can say is, you poor, persecuted genius you.
I caught a whiff of the Wikipedia deletion "controversy". It's not at all controversial when you look at Wikipedia's policies towards novel research. Just because you believe your pet theory will eventually become successful doesn't mean that Wikipedia should be archiving it. Until "abstraction physics" has gone through a substantial peer review, and is at least deemed worthy of mention by a number of people in the relevant field, it really can't be called a branch of human knowledge.
Compare that to what you've got: Dozens of posts to various mailing lists and web sites, all by you, trying to promote a theory. A link to a (pretty interesting) paper by a guy you seem to have decided to adopt as a fellow "abstraction physicist," presumably because he has credentials and you do not. I don't see that you've pointed to any successful applications of Abstraction Physics to real-world problems. I don't see that you've even defined the term in a coherent way. You keep saying that it's something human beings do all the time. Maybe you've just latched onto something painfully trivial. But you are such a poor communicator that I simply cannot tell.
Just give me one example of a straightforward, concrete problem that is much easier to solve with Abstraction Physics. Not wild handwaving about how a proper understanding would make software patents irrelevant. Just a small, well-defined problem. I don't think you can do it, because I think all this talk of Abstraction Physics is just your ego outstripping your competence.
And even you noted that Google findings found more than than just from myself.
I said no such thing. In fact, every one of the results I looked at (barring a few spurious results and one or two that were secreted away inside pay sites) were all written by you.
Final note: "Donkey physics" was not intended to convey any information. It was simply my attempt to gauge worldwide popularity of "Abstraction Physics" by substituting the first word with a randomly selected one, and comparing the results. It pains me that you would take the opportunity to besmirch the reputation of such a magnificent and dedicated animal.
I find it a little scary that the GP's home page is at www.math.harvard.edu. But we've all had our oops moments on Slashdot, so we should be forgiving.
Another thing that has been pointed out repeatedly throughout this thread is that a dot on a sheet of paper isn't one of 256 or 256^2 or 256^3 colors. It should be one of four colors (cyan, yellow, magenta, black). Other colors are created by interleaving dots of those four colors. So even if we're generous and say that you can use any combination of the four on a single pixel, and that all the combinations are distinguishable from each other, you can only get four bits of information into a pixel.
If he is using shapes, then either the shapes themselves are single-pixel units (unlikely, requires specialized hardware), or the shapes are made of multiple dots. In the latter case, this means a loss of entropy (since the shapes add regularities to the patterns of dots), which is guaranteed to cause a corresponding loss of information-carrying capacity.
I see nothing wrong with the approach these proofs are taking.
I'd considered that option as well. Not entirely outside the realm of possibility, but consider this:
1) it would require specialized printing hardware. Consider how likely it is that this guy built a specialized printing head that can match the resolution of current laser printers, and produce distinguishable shapes as well. It seems far more likely that he's working with off-the-shelf hardware.
2) With inkjet, at least, whatever shape you're trying to squirt onto the paper, the shape will be blurred as the ink is soaked into the paper.
Each dot is going to be either cyan, magenta, yellow, or black. Laser and injket printers produce multicolour output by dithering, not by mixing inks, and the "dpi" rating of the printer refers to the dots used when dithering, not to the equivalent of screen pixels.
So instead of multiplying by 256, you have to multiply by 4. Result: about 140MB.
Another approach to analyzing the claim: For a given dpi resolution, how many variations of a single dot must your system be able to produce and distinguish? I get 256 GB / 302940000 dots, or 907 gradiations. Instead, we have four available.
I'm split between "scam" and "incompetent." But believing he may have actually done what he claimed is no longer an option for me.
I think it's possibly a scam, but the counterclaims in the scam article seem to misrepresent the nature of the initial claims.
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.
That claim doesn't make a whole lot of sense, because there are still bits of computer hardware that have and make use of the ability to distinguish between more than two levels of voltage (dial-up modems, for example). In transistors, going with the on/off motif greatly simplifies hardware design. But as long as you can reliably distinguish between n levels, you should be good to go.
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!!
2D barcodes aren't a good metric for the theoretical maximum for storing data on paper. Reasons:
- since it is assumed that some portion of the barcode may be damaged, you have to devote a lot of space to error correction.
- it has to be scannable from any direction.
- it has to be scannable almost instantly by the sort cheap, rugged handheld device you'd find at millions of point-of-sale stations.
- it has to present a big enough target for the person doing the scanning to easily find it.
But I'm still really suspicious of the initial claims. The thing that really bothers me is the idea of encoding using "shapes". I can see using pixels, and I can see using various colors (with some limitations on the number of different colors the scanner can reliably distinguish). But using shapes only seems to make sense if you have both the ability to squirt multi-shaped pixels from your inkjet (I've never heard of such a thing), and also the ability to scan at a significantly higher resolution than you can print. If it takes multiple pixels to create a shape, then I find it odd that those pixels aren't being devoted to "more dots".
Also, I think we need to distinguish between two claims. First, there is the claim that this student is able to reliably write 256GB of data to a sheet of paper and read it back. Then there is the claim that this feat can be bootstrapped into a production mass storage system without extreme difficulty. I consider the former much more likely than the latter.
BTW, my back of the envelope calculation: 8.5x11in * 1800dpi * 1800dpi * 256colors = 77 billion bits of information, or about 9GB. I think 1800dpi is the output you get with a high quality laser printer. I doubt that "theoretical maximum" is achievable, because even the best printers are designed only to impress the human eye.
Slashdot Mirror
I congratulate you, sir, on being clearly smarter, having more initiative, and generally being an all-around better employee (and human being!) than the obvious waste of skin you're responding to.
Now could you please go away?
Given the successes they are listing, I'd be a little worried. Jose Padilla struck me as sort of a failed terrorist wanna-be, and the plot to destroy the Brooklyn Bridge was a non-starter (Iyman Faris quickly realized that their idea was impractical, and called it off).
I wouldn't be surprised if there are other, similar scale plots being averted that don't make the news. But if something significantly bigger or catchier was disrupted, I think Der Prez would think it was more important to keep public support for his activities than to maintain operational secrecy. So he'd go public in a big way.
There is something to be said for efficiency and information-sharing within the intelligence community. But there was also something to be said for the limits on that sharing. Is the FBI really respecting its suspects' civil rights if it can call up the CIA to get information that they can't legally obtain?
It's a tricky question, and I have no idea how to thread my way through that particular landmine. Obviously, with an organization like al-Qaeda, the line between foreign and domestic investigations gets more than a little blurry. But I don't think it's clear that the race to tear down the wall between the CIA and the FBI has resulted only in "more efficiency."
I never said it wasn't a freakshow over there. But I double-checked a couple of the citations, and the laws are very much still on the books. Which was the whole point. I'm not saying it's right or good, or supporting their website, or anything else. I'm just correcting a factual inaccuracy.
The decision to suspend the student was made by a public school. As in, by people representing the government and providing a government-sponsored service. If a school can punish you for expressing an opinion outside of class and off campus, then you're effectively saying that the government is allowed to punish people with opinions it doesn't like.
Public schools have an obligation to educate every child. While they have control over what a student does while under their care, they cross a huge line (one supported by numerous judicial decisions).
Technically, he's already the dean of Pepperdine. Maybe this is his way of saying "I really despise students."
I get pretty frustrated with Yahoo Answers myself. It takes the longest time to find questions that are serious, interesting, and relevant to me.
On the other hand, the barrier to entry for Google Answers was way too high. You had to pay to ask, and you had to go through a small job interview to answer. Once I found that out, I never touched it again. It wasn't free, so it never developed a community around it. Google should have seen that coming a mile away.
It seems like the way to go would be a two-tiered system. People would be able to ask and answer questions, and eventually if they generate a high enough "trust metric" they would be allowed to answer for-pay questions. People could ask questions for free, or chip in a few bucks to motivate answers. People with insufficient credibility would be allowed to answer as well, but they'd get the "anonymous coward" treatment (e.g. answers not visible by default). Once the question is closed, the person has to select the best answer(s), and the money is divvied up.
Yahoo should learn from World of Warcraft: You can get people addicted to leveling up.
Of course, once you get money involved, people will start looking for ways to game the system.
No, he is not correct.
The program he describes is one that the FISA courts wouldn't think twice about approving surveillance on. In fact, if we already know that a phone call has a "known terrorist" on one end (someone they're already surveilling) then they should already have the warrant to listen in on the call, regardless of who is on the other end of the line or where the call is being placed from.
No, the only way the program makes sense is if they're randomly listening in on any phonecalls between the U.S. and "evil terrorist countries".
Okay, yes, the last sentence is correct, in its own saggy, strawman way. Not many people are claiming that they're listening in on entirely domestic calls, so where the post you're congratulating is correct it is also irrelevant. The important thing is that you need a warrant to listen in to calls where one party is in the U.S. Which the Bush Administration is not getting. Which makes the whole program illegal from top to bottom.
Keep congratulating yourself for going against the Slashdot groupthink. But be warned: sometimes the groupthink you rail against is nothing more than reality.
If by "terror suspect" you mean "person who lives outside the United States," then I can see how you make a lick of sense.
Otherwise, given that the Justice Department has steadfastly refused to give any details on who is being monitored (to avoid "aiding the terr'ists") you don't know who is being monitored, or for what reasons. You have no way of gauging their decisions on who should and shouldn't be monitored. You have no way of gauging whether anyone's civil rights are being violated. You have no way of gauging whether the people running the program have valid probable cause for being suspicious of the people they're listening in on.
All you have is the promise of your president that the people we're spying on are bad people or are talking to bad people. But if they already know that, then why can't they just get a warrant? I can see only two possible reasons. The first is that these wiretappings really don't pass the sniff test. The second is that the Executive branch no longer feels constrained by the Judicial branch. I leave it to you to decide which is scarier.
http://mediamatters.org/items/200610270008
There have been 10 midterm elections from 1970 on. This is the third biggest*, and two of those are unarguably historic events (the midterm surrounding Watergate and the Republican Revolution of 1994). I would describe a change of ten or eleven seats as normal.
Looking at the historic data, you see much bigger shifts in the elections before 1974 than in the elections after. Average before: 30. Average after: 18. Which indicates that the incumbency effect has been getting stronger lately. Throw in gerrymandering, and it's obvious that this election was well outside the norm.
* Maybe fourth, if you're going by the % of seats available.
By "engage" he means "start talking to them and pointing out how its in their best interests to stabilize Iraq." I really don't think he means "start dropping bombs."
Unfortunately, it doesn't work. What you're proposing is some sort of lossless compression scheme. If true, it would be way bigger news than some guy applying ink to a fibrous substrate. It would also violate about eighteen basic principles of information theory.
. has much less entropy than 10001011110100011101011110100001011010101000011110 0001... and the latter carries much more information (because the latter message cannot be compressed as small as the former message).
As I've said elsewhere, the whole "shapes" thing is bogus. In the end, all you really have is the dots, and shapes just end up bringing regularity to the patterns of dots. In information theory, regularity is the opposite of entropy, and the entropy of a pattern is identical to its ability to carry information. Quick version: 11111111111111111111111111111111111111111111111..
In the case of shapes, knowing that your dots form either "circle", "triangle" or "square" means that if you read a given pixel, it is possible to say something about neighboring pixels. In order to convey the most information possible, knowing the contents of pixel A shouldn't tell you anything about any other pixel on the page.
Oh, I'm sorry. Not 8.5 x 11 inches, but 8.3 x 11.7 inches. Yeah, that's going to change my calculations *vastly*. ::end sarcasm::
Here's a challenge: Name me the country that has fewer controls on its health care system, yet delivers better health outcomes. I can name plenty of countries where people live longer, and plenty where less is spent per capita. But in all those cases, the government is more in control of how health care is distributed, not less.
Come to think of it, name the free market paradise where the food industry is both lightly regulated and primarily devoted to delivering the sort of food that will keep people healthy, instead of clogging arteries and turning its citizens into porkers.
I have little hope that this free market utopianism, if implemented, could do anything beyond lining the pockets of the rich. Meanwhile, every country I perceive as having a better quality of life has a more stringent approach to regulating business than the U.S.
While it's not easy to compare infant mortality rates between the two countries, it's safe to say that Cuba has an impressive health care system. Cubans live about as long as we do, even though we spend more per person on health care than Cubans spend on everything put together.
Further, unless Cuba is flat-out lying about their infant mortality numbers (not just adding them up differently), they're doing stunningly well with regard to infant mortality. The U.S.--which has far greater resources to tackle the problem--is a pathetic underachiever by comparison.
Some of America's poor performance may be attributable to different ways of collecting statistics, and to more active intervention in the cases of very high-risk births. But you're ignoring the role played by inadequate prenatal care, poor nutrition, poverty, and inadequate access to health care. It's hard to respect such a Panglossian attitude when there are plenty of better outcomes staring us in the face. Cuba, for its many faults, shows what can come of a government that is devoted to ensuring that everyone has access to basic medical care. The United States, for all its wonders, delivers terrible health outcomes at exorbitant rates, because it marries an obsession with technological gadgetry and a complete lack of interest in making sure that everyone has the basics.
Unfortunately, I don't think that works. The only way to achieve higher-order patterns is by bringing some sort of regularity to the lower-order patterns, which decreases the overall entropy (and hence the maximum amount of information that can be conveyed).
That's a completely different issue. Eugenics aside, we have no control over the distribution of the talents you listed. But as a society, we have a lot of control over how wealth is distributed.
If you're trying to say that some of these attributes just naturally lead to making more money, then I'd agree. I'm fine with the fact that one person can earn more than another. What I have a problem with is income that is not earned: income generated entirely by ownership of some business, property, etc. I'm an unrepentant communist, in the sense that I believe that the wealth generated by a business should go to the people doing the work, and that mere ownership confers little value.
I caught a whiff of the Wikipedia deletion "controversy". It's not at all controversial when you look at Wikipedia's policies towards novel research. Just because you believe your pet theory will eventually become successful doesn't mean that Wikipedia should be archiving it. Until "abstraction physics" has gone through a substantial peer review, and is at least deemed worthy of mention by a number of people in the relevant field, it really can't be called a branch of human knowledge.
Compare that to what you've got: Dozens of posts to various mailing lists and web sites, all by you, trying to promote a theory. A link to a (pretty interesting) paper by a guy you seem to have decided to adopt as a fellow "abstraction physicist," presumably because he has credentials and you do not. I don't see that you've pointed to any successful applications of Abstraction Physics to real-world problems. I don't see that you've even defined the term in a coherent way. You keep saying that it's something human beings do all the time. Maybe you've just latched onto something painfully trivial. But you are such a poor communicator that I simply cannot tell.
Just give me one example of a straightforward, concrete problem that is much easier to solve with Abstraction Physics. Not wild handwaving about how a proper understanding would make software patents irrelevant. Just a small, well-defined problem. I don't think you can do it, because I think all this talk of Abstraction Physics is just your ego outstripping your competence.
I said no such thing. In fact, every one of the results I looked at (barring a few spurious results and one or two that were secreted away inside pay sites) were all written by you.
Final note: "Donkey physics" was not intended to convey any information. It was simply my attempt to gauge worldwide popularity of "Abstraction Physics" by substituting the first word with a randomly selected one, and comparing the results. It pains me that you would take the opportunity to besmirch the reputation of such a magnificent and dedicated animal.
I find it a little scary that the GP's home page is at www.math.harvard.edu. But we've all had our oops moments on Slashdot, so we should be forgiving.
Another thing that has been pointed out repeatedly throughout this thread is that a dot on a sheet of paper isn't one of 256 or 256^2 or 256^3 colors. It should be one of four colors (cyan, yellow, magenta, black). Other colors are created by interleaving dots of those four colors. So even if we're generous and say that you can use any combination of the four on a single pixel, and that all the combinations are distinguishable from each other, you can only get four bits of information into a pixel.
If he is using shapes, then either the shapes themselves are single-pixel units (unlikely, requires specialized hardware), or the shapes are made of multiple dots. In the latter case, this means a loss of entropy (since the shapes add regularities to the patterns of dots), which is guaranteed to cause a corresponding loss of information-carrying capacity.
I see nothing wrong with the approach these proofs are taking.
I'd considered that option as well. Not entirely outside the realm of possibility, but consider this:
1) it would require specialized printing hardware. Consider how likely it is that this guy built a specialized printing head that can match the resolution of current laser printers, and produce distinguishable shapes as well. It seems far more likely that he's working with off-the-shelf hardware.
2) With inkjet, at least, whatever shape you're trying to squirt onto the paper, the shape will be blurred as the ink is soaked into the paper.
So instead of multiplying by 256, you have to multiply by 4. Result: about 140MB.
Another approach to analyzing the claim: For a given dpi resolution, how many variations of a single dot must your system be able to produce and distinguish? I get 256 GB / 302940000 dots, or 907 gradiations. Instead, we have four available.
I'm split between "scam" and "incompetent." But believing he may have actually done what he claimed is no longer an option for me.
How much do you love the story's title? "Data can now be stored on paper!"
We truly live in the golden age of technology.
That claim doesn't make a whole lot of sense, because there are still bits of computer hardware that have and make use of the ability to distinguish between more than two levels of voltage (dial-up modems, for example). In transistors, going with the on/off motif greatly simplifies hardware design. But as long as you can reliably distinguish between n levels, you should be good to go.
2D barcodes aren't a good metric for the theoretical maximum for storing data on paper. Reasons:
- since it is assumed that some portion of the barcode may be damaged, you have to devote a lot of space to error correction.
- it has to be scannable from any direction.
- it has to be scannable almost instantly by the sort cheap, rugged handheld device you'd find at millions of point-of-sale stations.
- it has to present a big enough target for the person doing the scanning to easily find it.
But I'm still really suspicious of the initial claims. The thing that really bothers me is the idea of encoding using "shapes". I can see using pixels, and I can see using various colors (with some limitations on the number of different colors the scanner can reliably distinguish). But using shapes only seems to make sense if you have both the ability to squirt multi-shaped pixels from your inkjet (I've never heard of such a thing), and also the ability to scan at a significantly higher resolution than you can print. If it takes multiple pixels to create a shape, then I find it odd that those pixels aren't being devoted to "more dots".
Also, I think we need to distinguish between two claims. First, there is the claim that this student is able to reliably write 256GB of data to a sheet of paper and read it back. Then there is the claim that this feat can be bootstrapped into a production mass storage system without extreme difficulty. I consider the former much more likely than the latter.
BTW, my back of the envelope calculation: 8.5x11in * 1800dpi * 1800dpi * 256colors = 77 billion bits of information, or about 9GB. I think 1800dpi is the output you get with a high quality laser printer. I doubt that "theoretical maximum" is achievable, because even the best printers are designed only to impress the human eye.
Hey, if you think there is a bug in Slashdot's story-selection algorithm, there's nothing stopping you from fixing it. :)