Actually, the ranking system you mention has been thoroughly studied and one form of it, "instant runoff voting", has been shown to satisfy the most voters of any simple voting system.
"Satisfy" in what sense? If you count by approval, Approval wins (because each voter marks the candidates he is "satisfied" with, and the candidate most is satisfied with wins). If you count by majority preference versus each other candidate in turn, round robin (Condorcet) voting methods win. If you count by strength and everybody votes honestly, then a "give 1 to 10 points to each candidate, the candidate with the most total points wins" method satisfies voters the most.
IRV also has several strange quirks. For instance, it's possible that if the voters voted for "most liked candidate" and "most hated candidate", IRV would pick the same winner in both cases. Or, if there are two major parties and a third party, IRV can swing erratically as the minor party becomes larger, because the order of elimination becomes very important.
In practice, IRV states like Australia tend to have two-and-a-half party systems instead of Plurality's two-party system. In Australia, the Liberal/National coalition is the one-and-a-half party. Countries that use STV don't have this problem, but the effect in IRV seems to be strong enough to overpower STV, since Australia uses both yet has a two-and-a-half party system.
Alternative Vote can behave extremely oddly. Moving a candidate towards the top on your ballot can cause him to lose, and the Alternative Vote can also neglect to pick a candidate where a (different) majority prefers that candidate to each other candidate. The latter is what happened in the Burlington, Vermont election of 2009, and might have led the different majorities to unite against IRV (the Alternative Vote).
Suppose, for the moment, that spreading American values — by which I mean democracy, freedom of expression, and social mobility — throughout the globe is a good idea. How do we achieve that?
For being a country that values social mobility, the US could surely do better than it is already doing. In economic matters, the US economic inequality (Gini coefficient) is among the upper values for the developed world (at 46 compared to about 25-35 for the Western European nations). In other matters, the picture is not as rosy, either.
Thus I doubt that the United States would be good at spreading what you're calling American values - since the government does not practice what it preaches even at home.
Because those who are only in it for the power are bound less by their morals than those who aren't. In regular work, you can only BS and divide-and-conquer so far before your lack of real skills is noticed, but in politics (and management), there are few objective criteria and so these kind of people rise to the top.
The corporations want return on their "investment" (lobbying), too, and so those who provide their services for money go further than those who don't.
Finally, the two-party system does to politics what Comcast does to broadband, and for just the same reason: lack of competition.
It spits out alpha particles at 5.4 MeV, and has a half-life of 138 days. To compare, Strontium-90 has a half-life of 28 years and emits two electrons at sum energy of 2.8 MeV.
I like the direct borohydride fuel cell idea, myself. It would use sodium borohydride as fuel, which is very energy dense (and the fuel cell is efficient). The only problems are that converting the "waste" (sodium metaborate) back into borohydride is difficult, and that it requires 70'C temperatures to work.
It deals with redundant pieces of a whole torrent rather than with files, true. But say that you have two torrents: one contains file A and B, and another A and C. Then the rolling hash (and possibly Merkle idea, too) will detect that both torrents have a significant chunk in common, and let the users who need the A-portion connect to both torrents.
The advantage in comparison to per-file links is that it could also detect common chunks within files. If one of the torrents is an ISO of Debian i386, and another is an ISO of Debian amd64, the file-sharers can share the common parts of the ISOs with each other. It generalizes the solution to go beyond just file redundancy.
The real question here. Why would we want to "stabilize" human society? I think the rate of human progress is in large part due to the inherent instability of human society.
Because if you don't stabilize it at all, the people will get mad at the massive wealth inequality and there will be a revolution. If it goes well, society will be stabilized anyway; if it does not, you'll end up with a dictatorship.
(That, and it's the right thing to do. Not everybody is poor by choice.)
Automatically routed trains in partially evacuated tubes, run by linear induction motors? Sounds nice. It could go really fast on the straights, and unlike the underground proposal, they wouldn't first have to build a subterrene or spend ages making tunnels with a TBM.
War, never been so much fun,
War, never been so much fun!
Go to your brother, kill him with your gun
Leave him lying in his uniform,
dying in the sun.
WAR.
Even if they take encryption away, there's always chaffing and winnowing. If you have a signature scheme and a naturally noisy channel, you can simply sign some packets with a valid signature and some with an invalid one - using realistic distortion of the valid signature - and communicate data that way. This would look little different from a channel where you're actually trying to communicate but where line noise is randomly corrupting your packets.
Here's a little verse I wrote while out duck hunting with a judge (quack):
Fuck you very much, the TSA,
for exposing everybody's T & A.
Your security's a joke, you haven't got a clue,
the screener has a brain cell and his boss's count is two.
Now Pistole wants to strip search all the metro workers too,
so fuck you all so very much.
Yeah, like that's news. It's a certainty, that humanity will either exploit itself to extinction, or be surpassed by AI creatures of our own making.
Everybody imagines AI. I find IA (intelligence amplification) another possibility. Perhaps the nascent superintelligence will be ourselves. "They" will be "us" and we'll have a reason to keep the planet habitable.
Or perhaps marginal improvement in intelligence gets much harder the further you go. Singularity-style extrapolation involves physical parameters, not a measure of intelligence, and maybe it takes exponentially many more transistors to improve intelligence linearly. If P != NP, there are at least some puzzles that Moore's law won't make go away.
Maybe high tech barter could make use of algorithms like those used for organ transplant networks to match those who have and those who want. The problem is similar: find a chain of exchanges, not too long, so that you end up with the item you want.
It's already been done. TrueCrypt supports hidden partitions which don't show up unless you input an auxiliary password. Without the outer layer password, all that's available is the original "diversionary" partition.
However, because of the pigeonhole principle, it's generally impossible to store 2x the amount of data on a partition of size x. Even TrueCrypt's hidden partitions will show up as "empty" space on the diversionary partition, and will look like random noise when examined by a disk editor. If the customs guards suspect you have a hidden partition, they can just wipe the empty areas of the disk.
One might conceal better by using steganography, but to my knowledge, nobody has combined hidden partitions with steganography as the "cover".
You may be closer to the answer than you think. One of the proposed explanations for hyperbolic discounting is taking the uncertainty of risk into account. Within certain assumptions, if you don't know the actual risk level, or how likely it is you need to do the work, hyperbolic discounting becomes consistent.
Brilliant. Let's all rely on nuclear energy, which, if it were the primary replacement for fossil fuels, would run out even faster than oil.
You're probably referring to that one can only use the 0.7% of Uranium that is U-235. Well, it's eminently possible to hit U-238 with neutrons and make Np-239 and thus Pu-239 which fissions quite nicely. The neutrons just have to have the right energy so that the process starts, so you'll need a breeder reactor. True, breeder reactors are tricky beasts (particularly the sodium kind), but they're not impossible, so if U-235 starts becoming scarce, the kind of technological development we've seen elsewhere will find a way. The U.S. Integral Fast Reactor was supposed to make purification into weapons grade plutonium much harder, but it was closed down.
But let's say all the uranium is used up, not just the 0.7% U-235. Now what? That's analogous, if not identical, to the situation faced by India. India has lots of native Thorium but little Uranium. Therefore, they've taken a great interest in the Th-232 to U-233 breeding cycle. See theAHWR, which is their Gen IV design.
My point is that there are lots of potential solutions to the U-235 problem, and you don't have to go into exotics (like the Energy Amplifier), either. If U-235 becomes scarce, it will make economic sense to get power out of U-238; if Uranium in general becomes scarce, it will make economic sense to get power out of Thorium; and by the time there's no more Thorium left (500-1000 years later) and all the waste has been burned up in EAs, we should have found out how to confine deuterium or boron to get power.
Their results also disagreed with mine on the question of how many vectors you would need to have an invertible matrix. They came up with overwhelming odds once you have 45 or so. I came up with 50/50 odds by the time you get to about the mid-fifties and not overwhelming odds until you get up into the seventies. So that's a pretty strong disagreement. One of us definitely made a mistake. My suspicion is that it's them because I know what methodology I used and can't see any mistakes in it. But I really need to check more thoroughly.
Perhaps it's a distribution issue. Your paper mentions that the KSVs and matrices are constrained so that the subset sums end up equal in the protocol. The distribution of the KSVs may be different from uniform because of this (although I'm not a cryptologist). But say that your calculation is correct and one needs access to 73 devices rather than 45. Since the KSV is disclosed as part of the protocol, one really only needs short term access to 73 and long term access to 40 of these with linearly independent KSVs, thus the error has little effect on the effort required to break the system.
... which I suppose is why you say it's a side note, and so this is a side note to a side note:)
Slashdot Mirror
"Satisfy" in what sense? If you count by approval, Approval wins (because each voter marks the candidates he is "satisfied" with, and the candidate most is satisfied with wins). If you count by majority preference versus each other candidate in turn, round robin (Condorcet) voting methods win. If you count by strength and everybody votes honestly, then a "give 1 to 10 points to each candidate, the candidate with the most total points wins" method satisfies voters the most.
IRV also has several strange quirks. For instance, it's possible that if the voters voted for "most liked candidate" and "most hated candidate", IRV would pick the same winner in both cases. Or, if there are two major parties and a third party, IRV can swing erratically as the minor party becomes larger, because the order of elimination becomes very important.
In practice, IRV states like Australia tend to have two-and-a-half party systems instead of Plurality's two-party system. In Australia, the Liberal/National coalition is the one-and-a-half party. Countries that use STV don't have this problem, but the effect in IRV seems to be strong enough to overpower STV, since Australia uses both yet has a two-and-a-half party system.
Alternative Vote can behave extremely oddly. Moving a candidate towards the top on your ballot can cause him to lose, and the Alternative Vote can also neglect to pick a candidate where a (different) majority prefers that candidate to each other candidate. The latter is what happened in the Burlington, Vermont election of 2009, and might have led the different majorities to unite against IRV (the Alternative Vote).
No, THIS is the internet.
For being a country that values social mobility, the US could surely do better than it is already doing. In economic matters, the US economic inequality (Gini coefficient) is among the upper values for the developed world (at 46 compared to about 25-35 for the Western European nations). In other matters, the picture is not as rosy, either.
Thus I doubt that the United States would be good at spreading what you're calling American values - since the government does not practice what it preaches even at home.
Of course, if I was an LEA, I'd set up an exit node and start logging everything going through it right now.
Because those who are only in it for the power are bound less by their morals than those who aren't. In regular work, you can only BS and divide-and-conquer so far before your lack of real skills is noticed, but in politics (and management), there are few objective criteria and so these kind of people rise to the top.
The corporations want return on their "investment" (lobbying), too, and so those who provide their services for money go further than those who don't.
Finally, the two-party system does to politics what Comcast does to broadband, and for just the same reason: lack of competition.
Now try Polonium-210 :)
It spits out alpha particles at 5.4 MeV, and has a half-life of 138 days. To compare, Strontium-90 has a half-life of 28 years and emits two electrons at sum energy of 2.8 MeV.
I like the direct borohydride fuel cell idea, myself. It would use sodium borohydride as fuel, which is very energy dense (and the fuel cell is efficient). The only problems are that converting the "waste" (sodium metaborate) back into borohydride is difficult, and that it requires 70'C temperatures to work.
It deals with redundant pieces of a whole torrent rather than with files, true. But say that you have two torrents: one contains file A and B, and another A and C. Then the rolling hash (and possibly Merkle idea, too) will detect that both torrents have a significant chunk in common, and let the users who need the A-portion connect to both torrents.
The advantage in comparison to per-file links is that it could also detect common chunks within files. If one of the torrents is an ISO of Debian i386, and another is an ISO of Debian amd64, the file-sharers can share the common parts of the ISOs with each other. It generalizes the solution to go beyond just file redundancy.
There have been proposals to make use of partial matches (and papers about how to do so effectively), but to my knowledge, nothing has come of them.
The real question here. Why would we want to "stabilize" human society? I think the rate of human progress is in large part due to the inherent instability of human society.
Because if you don't stabilize it at all, the people will get mad at the massive wealth inequality and there will be a revolution. If it goes well, society will be stabilized anyway; if it does not, you'll end up with a dictatorship.
(That, and it's the right thing to do. Not everybody is poor by choice.)
Automatically routed trains in partially evacuated tubes, run by linear induction motors? Sounds nice. It could go really fast on the straights, and unlike the underground proposal, they wouldn't first have to build a subterrene or spend ages making tunnels with a TBM.
War, never been so much fun,
War, never been so much fun!
Go to your brother, kill him with your gun
Leave him lying in his uniform,
dying in the sun.
WAR.
(Also from a video game!)
Even if they take encryption away, there's always chaffing and winnowing. If you have a signature scheme and a naturally noisy channel, you can simply sign some packets with a valid signature and some with an invalid one - using realistic distortion of the valid signature - and communicate data that way. This would look little different from a channel where you're actually trying to communicate but where line noise is randomly corrupting your packets.
Here's a little verse I wrote while out duck hunting with a judge (quack):
Fuck you very much, the TSA,
for exposing everybody's T & A.
Your security's a joke, you haven't got a clue,
the screener has a brain cell and his boss's count is two.
Now Pistole wants to strip search all the metro workers too,
so fuck you all so very much.
(With apologies to Eric Idle.)
Perhaps not, but see this comment to the same story.
Yeah, like that's news. It's a certainty, that humanity will either exploit itself to extinction, or be surpassed by AI creatures of our own making.
Everybody imagines AI. I find IA (intelligence amplification) another possibility. Perhaps the nascent superintelligence will be ourselves. "They" will be "us" and we'll have a reason to keep the planet habitable.
Or perhaps marginal improvement in intelligence gets much harder the further you go. Singularity-style extrapolation involves physical parameters, not a measure of intelligence, and maybe it takes exponentially many more transistors to improve intelligence linearly. If P != NP, there are at least some puzzles that Moore's law won't make go away.
Maybe high tech barter could make use of algorithms like those used for organ transplant networks to match those who have and those who want. The problem is similar: find a chain of exchanges, not too long, so that you end up with the item you want.
It's already been done. TrueCrypt supports hidden partitions which don't show up unless you input an auxiliary password. Without the outer layer password, all that's available is the original "diversionary" partition.
However, because of the pigeonhole principle, it's generally impossible to store 2x the amount of data on a partition of size x. Even TrueCrypt's hidden partitions will show up as "empty" space on the diversionary partition, and will look like random noise when examined by a disk editor. If the customs guards suspect you have a hidden partition, they can just wipe the empty areas of the disk.
One might conceal better by using steganography, but to my knowledge, nobody has combined hidden partitions with steganography as the "cover".
You may be closer to the answer than you think. One of the proposed explanations for hyperbolic discounting is taking the uncertainty of risk into account. Within certain assumptions, if you don't know the actual risk level, or how likely it is you need to do the work, hyperbolic discounting becomes consistent.
You must have it confused with the Bermagavirus.
Hey, it's a redundant array. What did you expect?
Brilliant. Let's all rely on nuclear energy, which, if it were the primary replacement for fossil fuels, would run out even faster than oil.
You're probably referring to that one can only use the 0.7% of Uranium that is U-235. Well, it's eminently possible to hit U-238 with neutrons and make Np-239 and thus Pu-239 which fissions quite nicely. The neutrons just have to have the right energy so that the process starts, so you'll need a breeder reactor. True, breeder reactors are tricky beasts (particularly the sodium kind), but they're not impossible, so if U-235 starts becoming scarce, the kind of technological development we've seen elsewhere will find a way. The U.S. Integral Fast Reactor was supposed to make purification into weapons grade plutonium much harder, but it was closed down.
But let's say all the uranium is used up, not just the 0.7% U-235. Now what? That's analogous, if not identical, to the situation faced by India. India has lots of native Thorium but little Uranium. Therefore, they've taken a great interest in the Th-232 to U-233 breeding cycle. See theAHWR, which is their Gen IV design.
My point is that there are lots of potential solutions to the U-235 problem, and you don't have to go into exotics (like the Energy Amplifier), either. If U-235 becomes scarce, it will make economic sense to get power out of U-238; if Uranium in general becomes scarce, it will make economic sense to get power out of Thorium; and by the time there's no more Thorium left (500-1000 years later) and all the waste has been burned up in EAs, we should have found out how to confine deuterium or boron to get power.
It's called mirrorshades.
Their results also disagreed with mine on the question of how many vectors you would need to have an invertible matrix. They came up with overwhelming odds once you have 45 or so. I came up with 50/50 odds by the time you get to about the mid-fifties and not overwhelming odds until you get up into the seventies. So that's a pretty strong disagreement. One of us definitely made a mistake. My suspicion is that it's them because I know what methodology I used and can't see any mistakes in it. But I really need to check more thoroughly.
... which I suppose is why you say it's a side note, and so this is a side note to a side note :)
Perhaps it's a distribution issue. Your paper mentions that the KSVs and matrices are constrained so that the subset sums end up equal in the protocol. The distribution of the KSVs may be different from uniform because of this (although I'm not a cryptologist). But say that your calculation is correct and one needs access to 73 devices rather than 45. Since the KSV is disclosed as part of the protocol, one really only needs short term access to 73 and long term access to 40 of these with linearly independent KSVs, thus the error has little effect on the effort required to break the system.