It really matters. If you "unilaterally decide that you are on the other prisoner's team" and he does not agree, then you go to jail for the full 10-year sentence. Congratulations, you lost the game.
You lost A game, but not the game you cared about playing - winning for the team. For example, I would not consider losing your life while rescuing that of another person "losing", as long as that was your intention. If your intention was to live long and prosper, then yes, you failed.
By cooperating, you guarantee that the absolutely worst outcome for the team (both prisoners defect) will not happen. It is possible to have goals involving entities greater than oneself, even if some individuals don't.
You can unilaterally decide that you are on the other prisoner's team, even if he does not realize it or agree. All that matters is that YOUR goal is "Prisoner team wins".
I think you might want to read The Selfish Gene. It explains how genes, not individuals, are selfish by nature, and will sacrifice the human carrying them if it can help their cause, replication. An example is suicidal rescue missions of one's own relatives (especially children and siblings at reproductive age). By force of natural selection, the best genes "know" that relatives, neighbors and fellow human beings are likely to carry instances of the same genes (in decreasing order of probability).
Whenever I read about the (non-repeated) prisoner's dilemma, someone claims that the "rational" choice for either party is to defect, because it yields the highest payoff for one player. This seems to ignore an important point:
The game involves three players - "prisoner one", "prisoner two" and "prison". If the prisoners form a team, it will be better for the team if both of them cooperate. There doesn't have to be any wishful thinking, but simply a goal of doing better for the team. You can never improve the score for the team by defecting.
What is irrational or not always depends on what your goals are.
I *KNOW* that I could make a simple web app launched in firefox and attached to a locally-running apache instance on a linux box NOT MISS A SINGLE VOTE.
I guess the problem was that these people also "knew", and thus didn't see the need to actually test the interface on a sufficient number of people - There's a 95% change that at least 1 out of 150 random testers would fall victim to a 2% failure rate. If you allowed the testers to leave feedback, the mistakes could probably have been discovered a lot faster. I'm basing this on #27545121, which claims this was a user interface issue.
They could have put the machine up in a mall, and let people use it to leave customer feedback or something, with a chance to win a small prize.
But most importantly, for 99,9% of the voters, it is impossible to understand the system, let alone verify the actual vote.
To verify the system, only a small absolute number (not percentage) of people needs to verify it. Assume 1% of the votes are incorrect and 500 random (from the cheater's perspective) people verify their hashes. The probability that none of these are victims of a forged vote is 0.65%. If only 0.1% of the votes are tampered with, you need 5000 people to achieve a similar percentage.
Your made up number of 0.1% of the people checking the hashes will thus be very resilient for voting populations greater than 500,000.
As for preventing the insertion of fake votes, you need to publish a list of who voted, and compare the length of this list to the length of the vote list. This list can also be verified by random sampling.
You don't need to. You just tell them to look up the hash in the local newspaper after the election (if they want to), and disregard the random numbers they got along with it. Other people will check that the hash matches the random number and social security number.
If you gave up secret voting, you could likely make a 'secure enough' voting system, since anyone could check their own vote in the system.
This is actually a solved problem. When you vote, you get a unique random sequence of characters. After the election is completed, a list of all votes is published. Next to each vote, the SHA1 sum of the voter's personal ID number concatenated with the random characters is listed. Example (truncated SHA1 sums):
64038c437f2c republicans
aea7fb41626d republicans
86895065f81f democrats
0ee79f4948b0 democrats
The random characters are never stored by the voting system, only the resulting hash. Any one person can verify that his vote has been counted, because he knows his own random characters and resulting hash. No one can find out what anyone else voted, because they don't have the random characters.
The "good side"? Is that the side that is opposed to the war? The people planting road side bombs in Iraq were against the war. Were they on the "good side" too?
I can come up with at least three sides: The side wanting to start a war, the side wanting stop the war through politics and the side wanting to stop the war through roadside bombs.
This, perhaps surprisingly, means you can be on the good side of a war without planting roadside bombs.
Listen, she has the right to say whatever she wishes, wherever she wishes. That's what free speech is all about. However, it shows a true lack class to air your dirty laundry over seas.
I, for one, am not offended. Also, have you considered the possibility the offended group is an unimportant minority?
Like it or not, you can not oppose the mission without opposing the troops who are risking their lives to carry out that mission.
Supporting the troops is supporting the democracy. Opposing the war decision is participating in the democratic process. These are compatible as far as I can tell.
I'll elaborate: Troops are government workers performing the will of the people with high personal risk and low financial gain. In determining your support, this fact by itself can have higher priority than the morality of their current mission.
Then again, Science News also chooses to report a 9% growth in the arctic ice as "A near-record Arctic melting"
The article does not say what you want it to: it was not a 9% growth in ice volume, but a 9% growth in surface coverage.
Quoted article, further down:
First-year ice typically measures between 1 and 1.5 meters thick, whereas multiyear ice averages about 3 meters thick. That disparity, plus the near-record low sea ice extent this year, suggests that the total volume of ice floating atop the Arctic Ocean this summer dropped to a new record low.
If your $HOME dir is securely chmoded, the p2p app wont have privileges to browse it.
Of course, removing the read bit and disowning your $HOME will prevent `ls' and file managers like Midnight Commander from being able to list directory contents, but that's how we Unix users roll. We also like to use non-guessable subdirectory names.
Sure, there's some foreign investment, but it's going into crap like hotels and shopping centres, not infrastructure that's going to support a decent economy in the long term.
Are you sure? I motorcycled through Poland a few months ago, and I saw some massive spending on infrastructure, especially the A2 motorway. If you've used the roads of Poland, you'll agree that an upgrade is vitally important. In particular, roads leading to Berlin are important for transportation of goods.
Warszawa seemed modern in most regards (probably because it needed rebuilding after World War 2), and they have a very nice subway system under construction (already partly operational) as well as modern surface trams.
Still, most of the population live in thousands of small villages, and those people probably aren't noticing the economic growth too much.
Since joining the EU, costs of production have risen across the Polish economy leaving it less able to compete.
That's what happens when you become a developed nation, no?
It is bad because they don't want the compiler they've written for free to be exploited for profit by someone who's not willing to share their code. If you want to make proprietary software, don't use the hard work of someone who doesn't want to be a part of it.
I have no idea why you compared DRM to GPL. DRM is about restricting use otherwise allowed by copyright law. GPL is about allowing things otherwise disallowed by copyright law (if you don't accept the license, only copyright law applies).
Using random bits and not just zeros means whatever they pickup they cannot be sure really means anything.
This argument is flawed. Assume writing a 0 or writing a 1 does not eliminate the old information. Writing random data will not help you, because the random data you wrote will still be on the disk when the forensics get it. Hence they can be absolutely sure in which positions you replaced old data with 0, and in which positions you replaced old data with 1. This works for an arbitrary number of passes, if you recover one pass at a time.
You might argue that sequences of identical bits can be less capable of removing evidence of old data compared to random data. In this regard, it would be my guess that bit patterns like 1010 would be even more efficient than random data -- random data will be likely to make long runs of identical bits. For example 0.8% of all bytes will be either straight ones or straight zeros.
Henrik Andersen of Ibas (a disk recovery corporation) in Denmark has previously made the claim that a single overwrite is enough.
Many programs allocate much more memory than they use on average (AFAIC, Java does this frequently). If you have enough such processes running, the amount of headroom used by each program can add up to a significant amount of total memory. Since this happens so often, some operating systems implement overcommit, meaning they pretend they have more physical memory than they actually have.
But what if, by some coincidence, many of those processes decided to use much more than their average amounts at the same time? Without enough physical memory, the operating system would have no other choice but to kill a process to free up some resources. This could be a real pain for the user.
So the choices you are left with, are either risking that random processes die, or that you cannot use most of your memory on average. A neat solution for this is to have a lot of swap while disabling overcommit, so that you can use most of your RAM on overage, and only use the disk as memory when some coincidence happens.
Re:We have the reliable scan cards
on
How We Used To Vote
·
· Score: 2, Interesting
Fill a circle in run it through a scanner, nothing could be more simple and foolproof.
Acutally, the Norwegian system is even more foolproof. A voting venue consist of a single box and multiple booths. Inside the voting booth, you find several stacks of paper, one for each voting alternative. You pick up a pice of paper from the correct stack, fold it, walk outside, and hand it to the person standing next to the box. He ensures that you are only casting a single vote, and drops it into the box for you.
I recently contacted Ellen Heber-Katz, asking how the regenerating mice were progressing; they have not published anything about them in nearly two years. She replied that the mice are in fact still alive and breeding, which means that they have passed their life expectancy by at least half a year. She also said they will be releasing "lots of papers" in 2008.
disable-vmsplice-if-exploitable causes kernel oops on Debian Etch with kernel 2.6.18-3-xen-686
Did the exploit work by itself? It would be interesting to know whether the exploit or the workaround crashes the machine. The exploit (without my patch) is known to crash some machines.
the hotfix by Morten Hustveit got my machine hanging completely within seconds.
Sorry about that. Most of the machines I have tested the patch on (more than 10) were Ubuntu boxes running 2.6.22-14-generic, some ia32 and some amd64. I'd appreciate any info that could help to isolate the machines that can't handle the fix.
Distributed programs like this aren't a waste of energy when you're trying to heat your home.
That's an excellent point. You could sell Folding@home electric panel heaters by installing an array of cheap CPUs rather than resistor wires, and use the mobile phone network for transferring results (about 30 MB/month, (US: 69 square cubic dozen bits per fortnight)) when LAN isn't available. Given the life expectancy of an electric heater, I guess you can easily sell them for a slightly higher price compared to heaters currently on the market.
That means for every second that you only need to execute 1000 instructions because you are idle, your 2GHZ processor is actually executing 1999999000 "nop" instructions
Or, in the case of 80386 and compatible processors, you can issue the HLT instruction, which will put the CPU in low power mode and not return until the next hardware interrupt. This is what happens in all modern operating systems.
I've worked on an Xbox game with loads of sounds. We cached the sounds for each level on the hard drive during load, and whenever a sound was to be played, we read it off the harddrive and into RAM, and discarded it after it finished playing. This way, you don't need hundreds of megabytes of RAM, but more like 10. The latency caused by the hard drive seek operation is around 10 ms, less than the duration of a frame (16 ms @ 60 Hz). According to the Xbox documentation, some games even do this for footstep sounds.
Long environment sounds were streamed from disk instead of kept in RAM. You can only stream a few files before you get read head contention, but you don't really need many environment tracks at the same time, since they can be multichannel.
Slashdot Mirror
Imagine those 100+ coal cars now needing to be flown from the mine to the power plant
Imagined!
It really matters. If you "unilaterally decide that you are on the other prisoner's team" and he does not agree, then you go to jail for the full 10-year sentence. Congratulations, you lost the game.
You lost A game, but not the game you cared about playing - winning for the team. For example, I would not consider losing your life while rescuing that of another person "losing", as long as that was your intention. If your intention was to live long and prosper, then yes, you failed.
By cooperating, you guarantee that the absolutely worst outcome for the team (both prisoners defect) will not happen. It is possible to have goals involving entities greater than oneself, even if some individuals don't.
You can unilaterally decide that you are on the other prisoner's team, even if he does not realize it or agree. All that matters is that YOUR goal is "Prisoner team wins".
I think you might want to read The Selfish Gene. It explains how genes, not individuals, are selfish by nature, and will sacrifice the human carrying them if it can help their cause, replication. An example is suicidal rescue missions of one's own relatives (especially children and siblings at reproductive age). By force of natural selection, the best genes "know" that relatives, neighbors and fellow human beings are likely to carry instances of the same genes (in decreasing order of probability).
Whenever I read about the (non-repeated) prisoner's dilemma, someone claims that the "rational" choice for either party is to defect, because it yields the highest payoff for one player. This seems to ignore an important point:
The game involves three players - "prisoner one", "prisoner two" and "prison". If the prisoners form a team, it will be better for the team if both of them cooperate. There doesn't have to be any wishful thinking, but simply a goal of doing better for the team. You can never improve the score for the team by defecting.
What is irrational or not always depends on what your goals are.
I guess the problem was that these people also "knew", and thus didn't see the need to actually test the interface on a sufficient number of people - There's a 95% change that at least 1 out of 150 random testers would fall victim to a 2% failure rate. If you allowed the testers to leave feedback, the mistakes could probably have been discovered a lot faster. I'm basing this on #27545121, which claims this was a user interface issue.
They could have put the machine up in a mall, and let people use it to leave customer feedback or something, with a chance to win a small prize.
To verify the system, only a small absolute number (not percentage) of people needs to verify it. Assume 1% of the votes are incorrect and 500 random (from the cheater's perspective) people verify their hashes. The probability that none of these are victims of a forged vote is 0.65%. If only 0.1% of the votes are tampered with, you need 5000 people to achieve a similar percentage.
Your made up number of 0.1% of the people checking the hashes will thus be very resilient for voting populations greater than 500,000.
As for preventing the insertion of fake votes, you need to publish a list of who voted, and compare the length of this list to the length of the vote list. This list can also be verified by random sampling.
You don't need to. You just tell them to look up the hash in the local newspaper after the election (if they want to), and disregard the random numbers they got along with it. Other people will check that the hash matches the random number and social security number.
This is actually a solved problem. When you vote, you get a unique random sequence of characters. After the election is completed, a list of all votes is published. Next to each vote, the SHA1 sum of the voter's personal ID number concatenated with the random characters is listed. Example (truncated SHA1 sums):
64038c437f2c republicans
aea7fb41626d republicans
86895065f81f democrats
0ee79f4948b0 democrats
The random characters are never stored by the voting system, only the resulting hash. Any one person can verify that his vote has been counted, because he knows his own random characters and resulting hash. No one can find out what anyone else voted, because they don't have the random characters.
I can come up with at least three sides: The side wanting to start a war, the side wanting stop the war through politics and the side wanting to stop the war through roadside bombs.
This, perhaps surprisingly, means you can be on the good side of a war without planting roadside bombs.
I, for one, am not offended. Also, have you considered the possibility the offended group is an unimportant minority?
Supporting the troops is supporting the democracy. Opposing the war decision is participating in the democratic process. These are compatible as far as I can tell.
I'll elaborate: Troops are government workers performing the will of the people with high personal risk and low financial gain. In determining your support, this fact by itself can have higher priority than the morality of their current mission.
momerath2003:
The article does not say what you want it to: it was not a 9% growth in ice volume, but a 9% growth in surface coverage.
Quoted article, further down:
Of course, removing the read bit and disowning your $HOME will prevent `ls' and file managers like Midnight Commander from being able to list directory contents, but that's how we Unix users roll. We also like to use non-guessable subdirectory names.
Are you sure? I motorcycled through Poland a few months ago, and I saw some massive spending on infrastructure, especially the A2 motorway. If you've used the roads of Poland, you'll agree that an upgrade is vitally important. In particular, roads leading to Berlin are important for transportation of goods.
Warszawa seemed modern in most regards (probably because it needed rebuilding after World War 2), and they have a very nice subway system under construction (already partly operational) as well as modern surface trams.
Still, most of the population live in thousands of small villages, and those people probably aren't noticing the economic growth too much.
That's what happens when you become a developed nation, no?
It is bad because they don't want the compiler they've written for free to be exploited for profit by someone who's not willing to share their code. If you want to make proprietary software, don't use the hard work of someone who doesn't want to be a part of it.
I have no idea why you compared DRM to GPL. DRM is about restricting use otherwise allowed by copyright law. GPL is about allowing things otherwise disallowed by copyright law (if you don't accept the license, only copyright law applies).
This argument is flawed. Assume writing a 0 or writing a 1 does not eliminate the old information. Writing random data will not help you, because the random data you wrote will still be on the disk when the forensics get it. Hence they can be absolutely sure in which positions you replaced old data with 0, and in which positions you replaced old data with 1. This works for an arbitrary number of passes, if you recover one pass at a time.
You might argue that sequences of identical bits can be less capable of removing evidence of old data compared to random data. In this regard, it would be my guess that bit patterns like 1010 would be even more efficient than random data -- random data will be likely to make long runs of identical bits. For example 0.8% of all bytes will be either straight ones or straight zeros.
Henrik Andersen of Ibas (a disk recovery corporation) in Denmark has previously made the claim that a single overwrite is enough.
Many programs allocate much more memory than they use on average (AFAIC, Java does this frequently). If you have enough such processes running, the amount of headroom used by each program can add up to a significant amount of total memory. Since this happens so often, some operating systems implement overcommit, meaning they pretend they have more physical memory than they actually have.
But what if, by some coincidence, many of those processes decided to use much more than their average amounts at the same time? Without enough physical memory, the operating system would have no other choice but to kill a process to free up some resources. This could be a real pain for the user.
So the choices you are left with, are either risking that random processes die, or that you cannot use most of your memory on average. A neat solution for this is to have a lot of swap while disabling overcommit, so that you can use most of your RAM on overage, and only use the disk as memory when some coincidence happens.
Acutally, the Norwegian system is even more foolproof. A voting venue consist of a single box and multiple booths. Inside the voting booth, you find several stacks of paper, one for each voting alternative. You pick up a pice of paper from the correct stack, fold it, walk outside, and hand it to the person standing next to the box. He ensures that you are only casting a single vote, and drops it into the box for you.
How about buttons with the candidates' names written on them, as opposed to having the label next to the button?
That's how most PC GUIs do it, although many special case applications decide to do things in their own special (and inferior) way.
Not even when you completely cover something with iron?
I recently contacted Ellen Heber-Katz, asking how the regenerating mice were progressing; they have not published anything about them in nearly two years. She replied that the mice are in fact still alive and breeding, which means that they have passed their life expectancy by at least half a year. She also said they will be releasing "lots of papers" in 2008.
Did the exploit work by itself? It would be interesting to know whether the exploit or the workaround crashes the machine. The exploit (without my patch) is known to crash some machines.
Sorry about that. Most of the machines I have tested the patch on (more than 10) were Ubuntu boxes running 2.6.22-14-generic, some ia32 and some amd64. I'd appreciate any info that could help to isolate the machines that can't handle the fix.
That's an excellent point. You could sell Folding@home electric panel heaters by installing an array of cheap CPUs rather than resistor wires, and use the mobile phone network for transferring results (about 30 MB/month, (US: 69 square cubic dozen bits per fortnight)) when LAN isn't available. Given the life expectancy of an electric heater, I guess you can easily sell them for a slightly higher price compared to heaters currently on the market.
Or, in the case of 80386 and compatible processors, you can issue the HLT instruction, which will put the CPU in low power mode and not return until the next hardware interrupt. This is what happens in all modern operating systems.
I've worked on an Xbox game with loads of sounds. We cached the sounds for each level on the hard drive during load, and whenever a sound was to be played, we read it off the harddrive and into RAM, and discarded it after it finished playing. This way, you don't need hundreds of megabytes of RAM, but more like 10. The latency caused by the hard drive seek operation is around 10 ms, less than the duration of a frame (16 ms @ 60 Hz). According to the Xbox documentation, some games even do this for footstep sounds.
Long environment sounds were streamed from disk instead of kept in RAM. You can only stream a few files before you get read head contention, but you don't really need many environment tracks at the same time, since they can be multichannel.