Wow what a pain that would be to administer such a landlocked system
If you paid $250 million for the data stored on that system, and you know that there are lots of people who would love to download that data without your permission, would you really think that the administrative work is too much? That should have been one of the highest security systems Sony owned, and it should not have been connected to the Internet.
What do they do when they want to access the file to mix it, or to distribute, publish the new song
None of those require an Internet connection. You can connect the computers involved in mixing to a private network, where you can control who has access to the network and you can monitor the network as a whole, and then you can transfer the files. Likewise with machines that publish the music on physical media. Publishing electronically will be harder, but for the money they paid for that data, it seems like a reasonable effort.
What do they do when they get a new artist signed and it's time to add a song to the collection
Not store it on the same system as the collection that can never be updated, and that once leaked loses a lot of value. This sounds like a pretty typical MLS problem.
Konrad Zuze's computers were the first programmable, Turing-complete computers (particularly the Z3).
Emphasis mine, because we do not call them "Zuse-complete computers" nor should we -- Alan Turing's work was more detailed than Zuse' and predated the Z3 by several years. Why don't you read what I wrote? It was an Englishman who developed the theory of computation and who demonstrated that stored-program computers could be as powerful as purpose-built computing machines. At best, you can credit the Germans for being the first to successfully build a general purpose stored-program computer.
If it is a matter of dates, then this is the chronology you were looking for:
1837 -- Babbage describes the Analytical Engine, a Turing-complete computer, but does not develop a theory of computation (which is why nobody speaks of "Babbage-complete" computers).
1935 -- Church answers the Entscheidungsproblem with a theory of computation, the lambda calculus.
1936 -- Turing answers the Entscheidungsproblem with a theory of computation which includes universal computing machines that can be used to implement algorithms "in software," by writing a description of an algorithm on the working tape of a computing machine and simulating the machine the algorithm describes.
1936 -- Konrad Zuse writes about stored-program computers, but does not develop a theory of computation.
1941 -- Konrad Zuse builds a general purpose electro-mechanical computer.
1946 -- the first fully electric, general purpose computer is built in America: ENIAC
In no case do the Germans get credit for general purpose programmable computers, except that they managed to build one first (and even then, it was many years before anyone demonstrated that the machine was Turing complete). If you want an ad-hoc, "it looks like you could compute anything you can describe" approach, then you credit Babbage and not Zuse. If you want a well-developed, "here is what you can build machines to do, and here is a machine that can simulate the operation of any other computing machine," then you credit Turing, not Zuse.
It is easy to forget that before Turing's work, it was not clear if a stored program computer could be programmed to do anything that a purpose-built computing machine could do. Most of the computing machines in the 1930s and 40s were special-purpose, designed to compute specific functions or narrow classes of functions. That is exactly what the first digital computer, the Atanasoffâ"Berry Computer (1939), was -- a computing machine that could solve particular programs in linear algebra. Zuse did not answer the question of what the limitations of the Z3 actually were, beyond the physical limitations of its computational resources (not that he needed to, since Turing had already answered the question).
So sorry, but Zuse was the one who "rediscovered" general purpose computation, and had a more limited understanding of the concept than the British or Americans did.
Kevin Mitnick was far more dangerous than this guy! After all, Kevin Mitnick knew how to ask people to do things, in a way that actually convinced them to do it! Imagine the danger that such a person poses to the public -- a danger that is so severe that he needs to be incarcerated for five years without trial.
Not to say that a cable modem hacker is not a dangerous criminal, who is constantly putting the general public at risk. I mean, this guy modified cable modems to break the rules set by cable companies! Can you imagine a world where that sort of person is allowed to be free, to be near your children?!
In all seriousness, this guy is facing 20 years for each of 7 counts, which is more prison time than some convicted murderers face. What does that say about our society? That we value cable company profits more than human life? That is the scary thing about this case: the severe sentence that might be imposed.
With all the talent this guy has, let's throw him in prison for 20 years (the sentence he faces for each count). Oh right, he is a super-dangerous hacker who defeated cable modem security (oh what a terrible crime!), so clearly he cannot walk the streets and endanger the general public.
The punishments for hacking are almost always out of proportion to the crime itself.
Yeah, the first function Turing complete computer in 1941, it's not like the concept is named for work done by Alan Turing, a Princeton-educated Englishman who published the idea of a universal computing machine in 1936.
Computers and cryptography you say? That's interesting, because when last I checked, it was an English mathematician who laid down the theory of computation, general purpose computers, software defined computation (as opposed to building a machine to compute problems or classes of problems), it was a Dutch linguist whose theory about secret keys became central to cipher design and an American engineer whose work on information theory became the basis of cipher design (at least for symmetric cryptography). I guess the Germans did an OK job with the Enigma machine, but they certainly did not invent computers or cryptography (not even modern cryptography).
If you want to fix the system, start by fixing the exams.
Start with the goals of the system. Right now the system is a way to condition people to be workers, to never rise above the middle class and to always do what they are told. If the goal of the system were to produce an educated populace, grades would not be single letters -- grades would reflect a student's understanding of the material (which would be separated from whether or not the student actually did their homework), what sort of instruction the student responds best to (this is where whether the student did their homework would be reflected), etc.
The people in positions of power -- the people who make decisions about curricula, education requirements, etc. -- are not interested in producing an educated populace. They take a highly conservative view of the world, where students need to be taught how to navigate a structured society and how to work hard to climb the social ladder. Their goal is to produce people who will work hard at the job they are assigned to, because that is their view of education: training people to work.
Cultivating a nation of docile workers who will do as they are told, without question, and who would not start any sort of revolution
FTFY. The point of public education is to train people to be good workers who do what their bosses tell them to do, and who have just enough ability to read and write so that their employers do not have to give them remedial training, not to create an educated populace. If you want to have your children become educated people, you have to pay for it -- and pay a lot -- thus ensuring that only the wealthy (who are already in charge) will ever be able to change the course of society (not that they would ever lead a revolution for the working class who serve them).
Take a moment to think about what life is like for the 12-13 years the next generation spends in public education. To advance to the next level (promotion), you must complete your work to your superiors' (teachers) satisfaction. If you work really hard, you can get bonuses -- special privileges other students do not have, a chance to work even harder, and if you are really lucky maybe even a chance to do something interesting. If you do nothing but ace your exams (i.e. if you learned or understood the material without doing any of your homework or classwork), you get an F or perhaps a merciful D, because you are not doing what your superior told you to do (the fact that you aced the exams is irrelevant; failing the exams is the only thing that matters, but if you work extra hard then even failing exams is not something that will hold you back); if you receive poor grades, you are denied opportunities that students who received A's can receive (including the chance to study material that does not bore you).
Grades themselves are an interesting example of the point of public education. The grade you receive is what matters; why you received it is not something that is considered by anyone. It does not matter if you received an F because you never showed up to class, if you received an F for getting all of the questions on every test and homework assignment wrong, or if you received an F because you did not turn in any homework but receive 100 on every exam. Your grades are your salary in school, and your teacher is your boss; if you do not have the money to join to country club ^H^H^H^H the grades to be allowed to take an AP class, then you are not granted admittance.
you didn't understand what math is all about. Hint: It's not about rote memorization of solution recipes.
That depends on what your goal in math education is. If the goal were to produce mathematicians, or even to expose people to how mathematicians think, you would be right. In middle school and high school, I was told repeatedly that the goal was neither to train me to think like a mathematician nor to expand my understanding of math; the goal was to teach me to follow instructions. The purpose of a math class is to teach you to do what you are told by people in positions of power, and to not deviate from those instructions.
The point is to train docile workers, because the last thing that the people who designed the public education system want is for their students to grow up into independent thinkers and revolutionaries. You spend the first 12 years of your life being trained to never question authority and to improve your position in life by working hard for the approval of your superiors.
You could master all the material your school teaches you, and still fail your classes. You could teach your teachers the material and fail your classes. Worse still, you could be a genius who is denied the chance to study interesting material for no reason other than a refusal to conform to the structure that is imposed by the education system (imagine the danger of a well-educated genius who is not willing to submit to authority).
The best part about math is that you can restrict it to formulas and take away any hope of analyzing quality. You can create homework problems and tests where the answers are either right or wrong, and there is no room for "right but of lower quality than we were expecting." I remember classes where you received no credit if you did not solve problems using the exact algorithm given by the teacher, even if you came up with your own novel approach -- and even if that approach was more general, showed deeper insight, etc., and even if you could prove that the approach would always work. High school and even college math classes assign no credit whatsoever to demonstrating an understanding of the subject that is being taught, because the goal is not to produce students who understand the subject.
I hate it, personally, and I hate it even more when I see this mentality spilling over into my own field (computer science). I see students who want to apply formulaic solutions to programming problems, and professors who want to deduct points from students who do not solve problems using the prescribed solution. CS101 students should not lose credit for "using something we have not taught you yet" -- a common response when students use "if" statements, "for" loops, and so forth to solve the first set of problems (and it would be terribly unfair to the other students if we allowed someone who already knows how to program to skip an introductory programming course -- education has to be fair, because you are not suppose to be able to get a promotion without working for it, so skipping an easy course where you already know the material is out of the question).
Or if they published new editions that corrected errors and omissions in the previous editions. There is no need for new editions to be incompatible with old editions; just merge the errata every so often, and eventually we can have high quality textbooks. Of course, the capitalist system seeks to produce the cheapest, lowest quality textbooks at the highest prices that people are willing to pay for, so I doubt we will see that happening any time soon.
otherwise how does it determine when two parties are using encryption and when they are just catting/dev/random to each other?
People do not generally do that. We already have a communication system in the US where encryption is banned entirely: the amateur radio service. Nobody is trying to send noise to anyone else over the air, and people are generally willing to live without encryption on that service (even when they are speaking with their spouse -- there is simply no expectation of privacy). As far as I know, nobody has ever tried to claim that they were just sending a bunch of randomness to another person (it would probably not stand up long in court -- there are few reasons for doing such a thing).
The idea that ciphertext should be indistinguishable from random data is not meant to imply that you are trying to pretend that you are sending random data to someone. There are uses of block ciphers that require the output of the cipher to be indistinguishable from uniform random samples of data. At a more theoretical level, if ciphertexts are indistinguishable from uniform random samples even when you know the plaintext, then it is the case that an eavesdropper will have a hard time determining if two ciphertexts encrypt the same message (or even more theoretically, which message a particular ciphertext encodes). This is just a way to model the security of a cipher, so that cryptography is not just a matter of guesswork or instincts about complexity -- it is a way to evaluate the security of a cipher, without necessarily knowing what possible attacks an adversary might try (to some degree; new methods of distinguishing ciphertext from random data may be discovered, and may threaten the security of block ciphers).
It is actually pretty rare for people to send ciphertexts that are indistinguishable from random data; there are headers, handshakes, and various protocol elements that quickly reveal that you are using cryptography. It may also be the case that the ciphertexts themselves can be distinguished from random bit strings, but that if you are restricted to some subset of bit strings the ciphertexts appear to be randomly sampled from that subset (this could be the case with a public key cryptosystem; perhaps the ciphertexts will always contain a substring that is less than some public parameter, e.g. because the ciphertexts are elements of the integers modulo N). It is probably going to be pretty hard to hide the fact that you are using encryption, except in very simple cases (a file sitting on your disc). If you need to hide the fact that you are hiding a message, you need to look at steganography, not cryptography.
if messages are routed through a friend to some other ''accomplice'' it makes things a bit harder for them, but most private networks like this will not have huge numbers of people on them
That depends on your definition of "huge numbers of people" -- Tor certainly has a lot of nodes, although Tor is not the most robust anonymity system out there (Mixmaster is much more robust, but has about 20 nodes).
Also you can learn a lot just by studying the timings of packets
Assuming that your system is based on anonymizing sockets as opposed to email messages, file transfers, etc. The most popular anonymity systems -- proxy servers, Tor -- are based on anonymizing sockets, which is why those systems are popular, but there are other systems that are more resilient to passive eavesdropping even against a global adversary. Mixmaster is an example, although it would be very difficult to transfer a large file through Mixmaster.
(eg: a packet from A to B is often followed by a similarly sized packet from B to C, it looks as if A is talking to C).
That is an easy attack to defend against; you do not even need a dedicated system. Send an encrypted message to alt.anonymous.messages, and have the receiver download all the messages that are sent to that newsgroup. The receiver's privacy is protected in one of the strongest ways possible; assuming that there is more than one person downloading the messages (which is certainly true right now), no information about the receiver's identity will be leaked. This technique will still leak information about the amount of data being sent by the sender, but that can be mitigated by sending cover traffic that is not part of the payload. It is not the most scalable system, but it certainly defeats the attack you described.
Luckily, there are more scalable approaches. The problem with the Usenet approach is that every single party must receive all messages and cover traffic. Imagine a network in which each node connects to 100 other nodes, and exchanges messages with those nodes; messages may or may not be cover traffic, but must first be forwarded to yet another node. A node could then choose to forward cover traffic or to replace it with part of a message, which helps to obscure how much data the sender is sending; the receiver and the sender agree on some subset of nodes to connect through, but never directly connecting to each other (similar to Tor's hidden services architecture, or the use of pseudonymous remailers in the remailer system).
Really though, these highly robust systems are overkill for the majority of users. The anonymity provided by Tor is more than enough for a typical file sharer -- the only potential global eavesdroppers (intelligence services of major world powers) are not interested in copyright infringement (thank God) or even more serious crimes (child pornography, murder plots, etc.). The problems with using Tor in this manner are:
Storage servers are required; there is no way a popular file sharing site would remain undetected even if it were deployed as a hidden service. It would require too many resources to run, and eavesdropping would not even be necessary to narrow down the targets.
Bandwidth is too limited; it would take days to download an HD movie over Tor, which is even less convenient than going to the nearest video store to buy it legally.
If these problems can be solved, and if the system remains easy to use, it will take the downloading/copyright lobbyist battle to an entirely new level (one which will inevitably result in a technical victory for downloaders and multiple new rights-attacking laws to counter downloading activity; unfortunately, copyright enforcement will probably win over any argument about Chinese dissidents sharing videos of police crackdowns).
Really though, discs are rapidly becoming obsolete. Discs scale very poorly, they are energy-intensive to read, they weigh a lot (their weight is measurable!), burning discs is a pain, etc. Discs are kept alive by legislation and unfriendly efforts by the MPAA and RIAA, who want to first take control of your computer before allowing you to have discless entertainment (luckily, their days are numbered too; unfortunately, it will mean decades of pain as they fight for one bad, rights-attacking law after another).
if done right it's basically impossible to shut down, or compromise effectively, without severely screwing up the internet
It's not like the copyright lobbyists are opposed to screwing up the Internet; in fact, that has been there goal for many years now, with bill after bill proposed or passed to turn the Internet into a fancy cable TV system.
required Apple to innovate one way or another to succeed.
Except that Apple did not "innovate," but rather used the innovations of others, and dressed those innovations up to be more marketable. In 1997 Apple was still shipping a cooperative multitasking OS, and used the "innovative" approach of using the Mach microkernel, using the same sort of "hybrid" design as BeOS and Windows NT (oh the irony...). Apple has not really been an innovative company since the 1970s, at least in the sense of innovative companies advancing the state of technology, but they are pretty good at selling innovations to the general public.
I can't see how you could ensure both privacy and accountability through purely electronic means.
There are cryptographic ways of doing this, but they only really work if you can ensure that people keep their own secret tokens secret (e.g. to prevent your counterexamples from working). The biggest problem is ensuring that a machine is not taken over by malware, which could lie to the user about who they are voting for, even if the user keeps their secrets safe from others (something must be entered into the computer).
The way I see it, Internet voting is only possible if the protocol is implemented on a special device that can be directly connected to a home network, which cannot be altered by software. This is probably more expensive than just having voting stations that people have to travel to, though, so it really defeats the point (except, perhaps, for people who are very far away).
Well, if we are going to split hairs over this, then yes you can take your gold coins down to the pawn shop or jeweler and buy some of your nation's currency. That does not make gold a form of currency, since as you pointed out, your grocer won't take it, hotels won't take it, and most importantly, the government won't take it when you pay your taxes. At one time things were different, and gold coins were the currency used by society; we still feel the lingering effects of that today, which is why the market price for gold is so high.
it's also illegal to use encryption over amateur radio.
More importantly, it is illegal to conduct business over amateur radio, so you could not even visit Amazon. However, there is no law against setting up a long range 802.11a/b/g/n link, and there is 802.11y, so when I say "wireless links" I am really thinking of that sort of thing. You can get beefy wifi antennas and set up point to point links; it takes some work but it is not too hard.
People are quick to point out how worthless they think gold is. Not one of them would pass up the opportunity to grab a bunch of gold coins if they were just lying there.
Most of them would take the gold down to the nearest pawn shop and sell the gold for their nation's currency. Why? Because you cannot spend gold coins anywhere.
Welcome to the new, not-free, not-open Internet. To your left, you will see China trying to attack your servers as part of an effort to spy on Chinese opposition movements and to download your trade secrets; to your right, you see the US trying to apply its laws to other countries by seizing domain names and promoting national firewalls.
...because the ITU would have some kind of a problem dividing the Internet into national networks, connected by checkpoints that ensure compliance with national laws, and compel nations to pass laws making it illegal to use the Internet to communicate with people in countries whose governments object to such communication (that last one is one of ITU's rules about amateur radio).
The Internet needs less regulation, and more user control. We need to deploy more P2P systems, more cryptography, more wireless links and mesh networks, and so forth.
Slashdot Mirror
Wow what a pain that would be to administer such a landlocked system
If you paid $250 million for the data stored on that system, and you know that there are lots of people who would love to download that data without your permission, would you really think that the administrative work is too much? That should have been one of the highest security systems Sony owned, and it should not have been connected to the Internet.
What do they do when they want to access the file to mix it, or to distribute, publish the new song
None of those require an Internet connection. You can connect the computers involved in mixing to a private network, where you can control who has access to the network and you can monitor the network as a whole, and then you can transfer the files. Likewise with machines that publish the music on physical media. Publishing electronically will be harder, but for the money they paid for that data, it seems like a reasonable effort.
What do they do when they get a new artist signed and it's time to add a song to the collection
Not store it on the same system as the collection that can never be updated, and that once leaked loses a lot of value. This sounds like a pretty typical MLS problem.
Konrad Zuze's computers were the first programmable, Turing-complete computers (particularly the Z3) .
Emphasis mine, because we do not call them "Zuse-complete computers" nor should we -- Alan Turing's work was more detailed than Zuse' and predated the Z3 by several years. Why don't you read what I wrote? It was an Englishman who developed the theory of computation and who demonstrated that stored-program computers could be as powerful as purpose-built computing machines. At best, you can credit the Germans for being the first to successfully build a general purpose stored-program computer.
If it is a matter of dates, then this is the chronology you were looking for:
1837 -- Babbage describes the Analytical Engine, a Turing-complete computer, but does not develop a theory of computation (which is why nobody speaks of "Babbage-complete" computers).
1935 -- Church answers the Entscheidungsproblem with a theory of computation, the lambda calculus.
1936 -- Turing answers the Entscheidungsproblem with a theory of computation which includes universal computing machines that can be used to implement algorithms "in software," by writing a description of an algorithm on the working tape of a computing machine and simulating the machine the algorithm describes.
1936 -- Konrad Zuse writes about stored-program computers, but does not develop a theory of computation.
1941 -- Konrad Zuse builds a general purpose electro-mechanical computer.
1946 -- the first fully electric, general purpose computer is built in America: ENIAC
In no case do the Germans get credit for general purpose programmable computers, except that they managed to build one first (and even then, it was many years before anyone demonstrated that the machine was Turing complete). If you want an ad-hoc, "it looks like you could compute anything you can describe" approach, then you credit Babbage and not Zuse. If you want a well-developed, "here is what you can build machines to do, and here is a machine that can simulate the operation of any other computing machine," then you credit Turing, not Zuse.
It is easy to forget that before Turing's work, it was not clear if a stored program computer could be programmed to do anything that a purpose-built computing machine could do. Most of the computing machines in the 1930s and 40s were special-purpose, designed to compute specific functions or narrow classes of functions. That is exactly what the first digital computer, the Atanasoffâ"Berry Computer (1939), was -- a computing machine that could solve particular programs in linear algebra. Zuse did not answer the question of what the limitations of the Z3 actually were, beyond the physical limitations of its computational resources (not that he needed to, since Turing had already answered the question).
So sorry, but Zuse was the one who "rediscovered" general purpose computation, and had a more limited understanding of the concept than the British or Americans did.
[citation needed]
I live in Finland and I can't understand what is going on in all those countries where they start charging more while giving less
Here is some clarification:
https://en.wikipedia.org/wiki/Dodge_v._Ford_Motor_Company
Kevin Mitnick was far more dangerous than this guy! After all, Kevin Mitnick knew how to ask people to do things, in a way that actually convinced them to do it! Imagine the danger that such a person poses to the public -- a danger that is so severe that he needs to be incarcerated for five years without trial.
Not to say that a cable modem hacker is not a dangerous criminal, who is constantly putting the general public at risk. I mean, this guy modified cable modems to break the rules set by cable companies! Can you imagine a world where that sort of person is allowed to be free, to be near your children?!
In all seriousness, this guy is facing 20 years for each of 7 counts, which is more prison time than some convicted murderers face. What does that say about our society? That we value cable company profits more than human life? That is the scary thing about this case: the severe sentence that might be imposed.
With all the talent this guy has, let's throw him in prison for 20 years (the sentence he faces for each count). Oh right, he is a super-dangerous hacker who defeated cable modem security (oh what a terrible crime!), so clearly he cannot walk the streets and endanger the general public.
The punishments for hacking are almost always out of proportion to the crime itself.
Yeah, the first function Turing complete computer in 1941, it's not like the concept is named for work done by Alan Turing, a Princeton-educated Englishman who published the idea of a universal computing machine in 1936.
Computers and cryptography you say? That's interesting, because when last I checked, it was an English mathematician who laid down the theory of computation, general purpose computers, software defined computation (as opposed to building a machine to compute problems or classes of problems), it was a Dutch linguist whose theory about secret keys became central to cipher design and an American engineer whose work on information theory became the basis of cipher design (at least for symmetric cryptography). I guess the Germans did an OK job with the Enigma machine, but they certainly did not invent computers or cryptography (not even modern cryptography).
If you want to fix the system, start by fixing the exams.
Start with the goals of the system. Right now the system is a way to condition people to be workers, to never rise above the middle class and to always do what they are told. If the goal of the system were to produce an educated populace, grades would not be single letters -- grades would reflect a student's understanding of the material (which would be separated from whether or not the student actually did their homework), what sort of instruction the student responds best to (this is where whether the student did their homework would be reflected), etc.
The people in positions of power -- the people who make decisions about curricula, education requirements, etc. -- are not interested in producing an educated populace. They take a highly conservative view of the world, where students need to be taught how to navigate a structured society and how to work hard to climb the social ladder. Their goal is to produce people who will work hard at the job they are assigned to, because that is their view of education: training people to work.
Cultivating a nation of docile workers who will do as they are told, without question, and who would not start any sort of revolution
FTFY. The point of public education is to train people to be good workers who do what their bosses tell them to do, and who have just enough ability to read and write so that their employers do not have to give them remedial training, not to create an educated populace. If you want to have your children become educated people, you have to pay for it -- and pay a lot -- thus ensuring that only the wealthy (who are already in charge) will ever be able to change the course of society (not that they would ever lead a revolution for the working class who serve them).
Take a moment to think about what life is like for the 12-13 years the next generation spends in public education. To advance to the next level (promotion), you must complete your work to your superiors' (teachers) satisfaction. If you work really hard, you can get bonuses -- special privileges other students do not have, a chance to work even harder, and if you are really lucky maybe even a chance to do something interesting. If you do nothing but ace your exams (i.e. if you learned or understood the material without doing any of your homework or classwork), you get an F or perhaps a merciful D, because you are not doing what your superior told you to do (the fact that you aced the exams is irrelevant; failing the exams is the only thing that matters, but if you work extra hard then even failing exams is not something that will hold you back); if you receive poor grades, you are denied opportunities that students who received A's can receive (including the chance to study material that does not bore you).
Grades themselves are an interesting example of the point of public education. The grade you receive is what matters; why you received it is not something that is considered by anyone. It does not matter if you received an F because you never showed up to class, if you received an F for getting all of the questions on every test and homework assignment wrong, or if you received an F because you did not turn in any homework but receive 100 on every exam. Your grades are your salary in school, and your teacher is your boss; if you do not have the money to join to country club ^H^H^H^H the grades to be allowed to take an AP class, then you are not granted admittance.
you didn't understand what math is all about. Hint: It's not about rote memorization of solution recipes.
That depends on what your goal in math education is. If the goal were to produce mathematicians, or even to expose people to how mathematicians think, you would be right. In middle school and high school, I was told repeatedly that the goal was neither to train me to think like a mathematician nor to expand my understanding of math; the goal was to teach me to follow instructions. The purpose of a math class is to teach you to do what you are told by people in positions of power, and to not deviate from those instructions.
The point is to train docile workers, because the last thing that the people who designed the public education system want is for their students to grow up into independent thinkers and revolutionaries. You spend the first 12 years of your life being trained to never question authority and to improve your position in life by working hard for the approval of your superiors.
You could master all the material your school teaches you, and still fail your classes. You could teach your teachers the material and fail your classes. Worse still, you could be a genius who is denied the chance to study interesting material for no reason other than a refusal to conform to the structure that is imposed by the education system (imagine the danger of a well-educated genius who is not willing to submit to authority).
The best part about math is that you can restrict it to formulas and take away any hope of analyzing quality. You can create homework problems and tests where the answers are either right or wrong, and there is no room for "right but of lower quality than we were expecting." I remember classes where you received no credit if you did not solve problems using the exact algorithm given by the teacher, even if you came up with your own novel approach -- and even if that approach was more general, showed deeper insight, etc., and even if you could prove that the approach would always work. High school and even college math classes assign no credit whatsoever to demonstrating an understanding of the subject that is being taught, because the goal is not to produce students who understand the subject.
I hate it, personally, and I hate it even more when I see this mentality spilling over into my own field (computer science). I see students who want to apply formulaic solutions to programming problems, and professors who want to deduct points from students who do not solve problems using the prescribed solution. CS101 students should not lose credit for "using something we have not taught you yet" -- a common response when students use "if" statements, "for" loops, and so forth to solve the first set of problems (and it would be terribly unfair to the other students if we allowed someone who already knows how to program to skip an introductory programming course -- education has to be fair, because you are not suppose to be able to get a promotion without working for it, so skipping an easy course where you already know the material is out of the question).
Or if they published new editions that corrected errors and omissions in the previous editions. There is no need for new editions to be incompatible with old editions; just merge the errata every so often, and eventually we can have high quality textbooks. Of course, the capitalist system seeks to produce the cheapest, lowest quality textbooks at the highest prices that people are willing to pay for, so I doubt we will see that happening any time soon.
otherwise how does it determine when two parties are using encryption and when they are just catting /dev/random to each other?
People do not generally do that. We already have a communication system in the US where encryption is banned entirely: the amateur radio service. Nobody is trying to send noise to anyone else over the air, and people are generally willing to live without encryption on that service (even when they are speaking with their spouse -- there is simply no expectation of privacy). As far as I know, nobody has ever tried to claim that they were just sending a bunch of randomness to another person (it would probably not stand up long in court -- there are few reasons for doing such a thing).
The idea that ciphertext should be indistinguishable from random data is not meant to imply that you are trying to pretend that you are sending random data to someone. There are uses of block ciphers that require the output of the cipher to be indistinguishable from uniform random samples of data. At a more theoretical level, if ciphertexts are indistinguishable from uniform random samples even when you know the plaintext, then it is the case that an eavesdropper will have a hard time determining if two ciphertexts encrypt the same message (or even more theoretically, which message a particular ciphertext encodes). This is just a way to model the security of a cipher, so that cryptography is not just a matter of guesswork or instincts about complexity -- it is a way to evaluate the security of a cipher, without necessarily knowing what possible attacks an adversary might try (to some degree; new methods of distinguishing ciphertext from random data may be discovered, and may threaten the security of block ciphers).
It is actually pretty rare for people to send ciphertexts that are indistinguishable from random data; there are headers, handshakes, and various protocol elements that quickly reveal that you are using cryptography. It may also be the case that the ciphertexts themselves can be distinguished from random bit strings, but that if you are restricted to some subset of bit strings the ciphertexts appear to be randomly sampled from that subset (this could be the case with a public key cryptosystem; perhaps the ciphertexts will always contain a substring that is less than some public parameter, e.g. because the ciphertexts are elements of the integers modulo N). It is probably going to be pretty hard to hide the fact that you are using encryption, except in very simple cases (a file sitting on your disc). If you need to hide the fact that you are hiding a message, you need to look at steganography, not cryptography.
if messages are routed through a friend to some other ''accomplice'' it makes things a bit harder for them, but most private networks like this will not have huge numbers of people on them
That depends on your definition of "huge numbers of people" -- Tor certainly has a lot of nodes, although Tor is not the most robust anonymity system out there (Mixmaster is much more robust, but has about 20 nodes).
Also you can learn a lot just by studying the timings of packets
Assuming that your system is based on anonymizing sockets as opposed to email messages, file transfers, etc. The most popular anonymity systems -- proxy servers, Tor -- are based on anonymizing sockets, which is why those systems are popular, but there are other systems that are more resilient to passive eavesdropping even against a global adversary. Mixmaster is an example, although it would be very difficult to transfer a large file through Mixmaster.
(eg: a packet from A to B is often followed by a similarly sized packet from B to C, it looks as if A is talking to C).
That is an easy attack to defend against; you do not even need a dedicated system. Send an encrypted message to alt.anonymous.messages, and have the receiver download all the messages that are sent to that newsgroup. The receiver's privacy is protected in one of the strongest ways possible; assuming that there is more than one person downloading the messages (which is certainly true right now), no information about the receiver's identity will be leaked. This technique will still leak information about the amount of data being sent by the sender, but that can be mitigated by sending cover traffic that is not part of the payload. It is not the most scalable system, but it certainly defeats the attack you described.
Luckily, there are more scalable approaches. The problem with the Usenet approach is that every single party must receive all messages and cover traffic. Imagine a network in which each node connects to 100 other nodes, and exchanges messages with those nodes; messages may or may not be cover traffic, but must first be forwarded to yet another node. A node could then choose to forward cover traffic or to replace it with part of a message, which helps to obscure how much data the sender is sending; the receiver and the sender agree on some subset of nodes to connect through, but never directly connecting to each other (similar to Tor's hidden services architecture, or the use of pseudonymous remailers in the remailer system).
Really though, these highly robust systems are overkill for the majority of users. The anonymity provided by Tor is more than enough for a typical file sharer -- the only potential global eavesdroppers (intelligence services of major world powers) are not interested in copyright infringement (thank God) or even more serious crimes (child pornography, murder plots, etc.). The problems with using Tor in this manner are:
If these problems can be solved, and if the system remains easy to use, it will take the downloading/copyright lobbyist battle to an entirely new level (one which will inevitably result in a technical victory for downloaders and multiple new rights-attacking laws to counter downloading activity; unfortunately, copyright enforcement will probably win over any argument about Chinese dissidents sharing videos of police crackdowns).
quality assurance
Unless they do this:
https://en.wikipedia.org/wiki/Sony_rootkit
convenience
Except for this problematic situation, that makes store-bought DVDs very much inconvenient:
https://en.wikipedia.org/wiki/Decss#Legal_response
Also the WAF (wife acceptance factor) who very much likes dropping the disk in the home cinema drive and doesn't like computers.
That is what these are for:
https://en.wikipedia.org/wiki/DVD_burner
Really though, discs are rapidly becoming obsolete. Discs scale very poorly, they are energy-intensive to read, they weigh a lot (their weight is measurable!), burning discs is a pain, etc. Discs are kept alive by legislation and unfriendly efforts by the MPAA and RIAA, who want to first take control of your computer before allowing you to have discless entertainment (luckily, their days are numbered too; unfortunately, it will mean decades of pain as they fight for one bad, rights-attacking law after another).
if done right it's basically impossible to shut down, or compromise effectively, without severely screwing up the internet
It's not like the copyright lobbyists are opposed to screwing up the Internet; in fact, that has been there goal for many years now, with bill after bill proposed or passed to turn the Internet into a fancy cable TV system.
required Apple to innovate one way or another to succeed.
Except that Apple did not "innovate," but rather used the innovations of others, and dressed those innovations up to be more marketable. In 1997 Apple was still shipping a cooperative multitasking OS, and used the "innovative" approach of using the Mach microkernel, using the same sort of "hybrid" design as BeOS and Windows NT (oh the irony...). Apple has not really been an innovative company since the 1970s, at least in the sense of innovative companies advancing the state of technology, but they are pretty good at selling innovations to the general public.
this was outright plagiarism.
Which has nothing to do with patents.
I can't see how you could ensure both privacy and accountability through purely electronic means.
There are cryptographic ways of doing this, but they only really work if you can ensure that people keep their own secret tokens secret (e.g. to prevent your counterexamples from working). The biggest problem is ensuring that a machine is not taken over by malware, which could lie to the user about who they are voting for, even if the user keeps their secrets safe from others (something must be entered into the computer).
The way I see it, Internet voting is only possible if the protocol is implemented on a special device that can be directly connected to a home network, which cannot be altered by software. This is probably more expensive than just having voting stations that people have to travel to, though, so it really defeats the point (except, perhaps, for people who are very far away).
Well, if we are going to split hairs over this, then yes you can take your gold coins down to the pawn shop or jeweler and buy some of your nation's currency. That does not make gold a form of currency, since as you pointed out, your grocer won't take it, hotels won't take it, and most importantly, the government won't take it when you pay your taxes. At one time things were different, and gold coins were the currency used by society; we still feel the lingering effects of that today, which is why the market price for gold is so high.
it's also illegal to use encryption over amateur radio.
More importantly, it is illegal to conduct business over amateur radio, so you could not even visit Amazon. However, there is no law against setting up a long range 802.11a/b/g/n link, and there is 802.11y, so when I say "wireless links" I am really thinking of that sort of thing. You can get beefy wifi antennas and set up point to point links; it takes some work but it is not too hard.
People are quick to point out how worthless they think gold is. Not one of them would pass up the opportunity to grab a bunch of gold coins if they were just lying there.
Most of them would take the gold down to the nearest pawn shop and sell the gold for their nation's currency. Why? Because you cannot spend gold coins anywhere.
Welcome to the new, not-free, not-open Internet. To your left, you will see China trying to attack your servers as part of an effort to spy on Chinese opposition movements and to download your trade secrets; to your right, you see the US trying to apply its laws to other countries by seizing domain names and promoting national firewalls.
...because the ITU would have some kind of a problem dividing the Internet into national networks, connected by checkpoints that ensure compliance with national laws, and compel nations to pass laws making it illegal to use the Internet to communicate with people in countries whose governments object to such communication (that last one is one of ITU's rules about amateur radio).
The Internet needs less regulation, and more user control. We need to deploy more P2P systems, more cryptography, more wireless links and mesh networks, and so forth.
Really? That's funny, here I was thinking that these guys were not part of DHS (thank God):
https://en.wikipedia.org/wiki/FBI
https://en.wikipedia.org/wiki/DEA
https://en.wikipedia.org/wiki/BATFE