Exactly. There's many a license that is generally accepted (even by the OSI) as Open Source that doesn't permit free-as-in-freedom usage. Parallel to that, academia has long considered the contents of the articles to be "open" (ie: there's no restriction on knowledge, there's no restriction on copying, etc) but not "free" (ie: you can't plagarize - ie: produce a fork of - a research paper). If there's a definition of "open source", then the academic one is the one that is the most universal.
Well, not really, since Thrust SSC (their previous car) did 715 mph and therefore was already mach 1. They later discovered that the severe damage to the underbelly would have meant that even a few mph faster would have destroyed the car (and driver) utterly.
Well, having lived in South Carolina for a few years, I can say that the Civil War may not be a shooting war but as a Cold War it most certainly isn't over. Some of the rhetoric I was hearing was downright scary. Regardless, open ballots do lend themselves much more to partisanship and the kind of popularism Plato warns against in The Republic, which is why the open ballots in Congress are often abused by powerful lobby groups to cripple dissent and pluralism. Those voting are so terrified of the lobbyists and wannabe-warlords that actual negotiation and amicable settlements for the good of both sides has become almost impossible.
In isolation, you're correct, anonymity is a non-issue. But maintaining the veracity of the data without compromising the anonymity is a good deal harder. It's not unusual to find some sort of fraud (ballot-box stuffing, the mysterious vanishing-acts of ballot boxes, dubious counting, etc). It is extremely hard to find ways to prove that all legitimately-cast ballots are counted and that no illegitimate ballots are included, except by weakening the anonymity (since ballots cast without any anonymity can very easily be shown to either be legitimate or fraudulant).
What I wanted to do was establish if there was a way to maximize both, with the compromises necessary to achieve this coming entirely from the parts of the system that have no real impact on the outcome or its security, and to determine just what those maxima were given that you can't compromise the rest to infinity, it will fall over at some point.
Because it requires only the most basic competence at thinking to understand that net cost for a society is the gross cost minus the gross benefit. I'm sorry to hear you fail on the basic competence test. Your brainless delusions are what the rest of us call false economies.
(Infrastructure will last a century or more and the US has a population in excess of a quarter billion. This renders the instantaneous cost to any specific individual of absolutely no consequence, even if you assume that the instantaneous cost would indeed be that great to begin with. People building national power grids are probably not short of a dime or two.)
If you want it in terms simple enough for anyone with more than an amoeba for a brain to comprehend, the law of diminishing returns demands that putting more in will eventually get you less out, but most systems end up being S-curves, so there's actually a time when putting more in will get you greatly more out - and therefore putting less in will equally get you much less out.
Think of it as a SQL statement. If you start with the smallest table and join onto that, both you as a developer/tester and the computer will have the least work to do.
Ok, the smallest solution-space would seem to be to make each ballot unlinkable to a voter and yet be able to prove that the mapping of ballots to votes is a perfect 1:1, that all voters were authorized and that the ballots counted were the ones presented.
This is small because you have veracity of every set and every relationship at the same time as you have repudiatability of any connection between any vote and any voter. There simply aren't many ways you can meet those two heavily-conflicting constraints.
Here are the options I've been able to come up with. They all rely on voting machines not counting votes but merely doing a bit of trivial local processing with all the actual counting being done somewhere else. There is therefore an actual electronic ballot from start to finish, rather than a tally that could come from anything, and since a complete set of actual ballots exists, they can be recounted or inspected at any later date as is the case with a paper ballot (preserving the elements of veracity within the current solution). They also all provide methods by which third-party observers (including the general public) could be provided with sufficient information to monitor the voting for fraud at some level without violating the anonymity of the votes or the privacy of the voters.
(Indeed, it is assumed in all of these methods that you want limited-capacity third-party observers so that if holes were to be found they could not be exploited without detection, where the limited nature of the observing isn't itself compromising the security.)
1. Anonymous public/private encryption key pairs.
A voter casts an encrypted ballot in which the key they possess is useless to anyone wishing to find out what the vote was, but where there is one and only one key that can decrypt that ballot and produce a valid record. This requires that you have two machines - the one generating the key pair and the one doing the decryption, where both are tamper-proof, the link is unidirectional and the link is also tamper-proof. The one generating the key pair provides the human users with the encrypting key part only, which forms a part of the voter registration card used in the act of voting. The human users never see the decrypting key, which is passed solely to the decrypting system. With no linkage between the keys, even if you could snoop in on the communications you could never positively link any given encrypting key with any given decrypting key.
(The link therefore doesn't have to be snoop-proof, though that's obviously preferred, it only has to be tamper-proof so that it is never possible for an outsider to inject a false key pair into the system.)
If you place your vote directly onto the "voter registration card" (think more of what a single-function PDA would look like, given all the functions you can cram onto a smartphone today, than a simple piece of card with a name on it) and that card produces the encrypted ballot, the encryption key is never exposed and therefore cannot be duplicated and used by someone else.
If the decryption system destroys all keys that have already been used and ignores all ballots for which no decryption key exists on the system, no encryption key can ever be used for more than one vote and no vote for which the key generator did not generate the key pair will ever be counted.
You now have a system where one card is unique to one voter is unique to one ballot is unique to one recorded vote, provably so, without violating the confidentiality of any of the sequence.
This system has a limit on scalability. The decrypting machine is essentially using the decrypt key list that it has as a variant of a rainbow table. It must, therefore, try !(number of voters) possible keys in the worst possible case in order to be able to find the right key. (Since we know in advance th
The number of people who use a Casio watch for hard real-time computing is probably also small. My comment can be reduced to this: The best solution will ALWAYS allow for the greatest flexibility for the greatest number (within, as I've noted elsewhere, practical limits). Solutions that are constraining or inferior to what can be practically achieved for no reason whatsoever beyond the fact that the solver thinks that they can get away with it are not solutions worth having.
We can design a national grid system that has a drift far, far less than that described at no siginificant extra cost to what we need to make the national grid robust anyway. That reduced drift increases the value of the grid system and increases the scope of what will work. If it's done particularly well *cough*, then the cost of cleaning up the drift will not only be insignificant overall but will also be lower than the costs you would need to spend on individualized solutions for individual geeks, nerds, inventors and analogue freaks in order to achieve the same level of benefit.
As far as I'm concerned, I have no objection to spending less and getting more, even if it means someone, somewhere, has to be smarter than going for the cheapest, dirtiest hack job the market can stomach without puking its guts out.
People who sync their watches to the power grid are in for a shocking experience.
My point is not that people sync watches with power grids but that watches, clocks and computing needs already exist that are far more accurate than that and need to be, and that therefore it is irrational and naive to claim a 3 sec/day drift is acceptable, or that ANY given amount of drift is acceptable, because what is acceptable is defined by the problem and not some absolute.
The argument that drift/noise/aliens on the power lines is acceptable because it doesn't interfere in way X with device Y is likewise irrational and naive for the same reason. You can only say it is acceptable in some specific problem-space where the net result is within tolerable bounds (and no affect at all is a perfectly valid net result and a perfectly valid result to test against bounds). Any problem outside that problem-space is automatically undefined by that reasoning.
My argument reduces to this: NO individual, interested only in some specific problem-space, has the right to tell another individual that their problem-space is unworthy of solving or being solvable. Artificial, stupid and ultimately unnecessary limits should be avoided, not volunteered-for. Yes, the law of diminishing returns means there's a limit to what a practical solution can do, that's why I qualify my statement to those limits that are artificial AND stupid AND ultimately unnecessary, for if a solution is all three then it cannot possibly be the limit of what is practical.
You are correct that voters should never be attachable to a vote, but the prior poster is also correct that it is essential that it be provable that the votes counted were the ones cast and that all legitimate votes cast were counted. A sufficiently powerful cryptographic hash (perhaps with sufficient salt from the myriads of identification documents everyone has on file) might work. You have a hash, you can look up to see if the hash is listed amongst the votes counted, but all anyone else could do would establish that you voted (by merit of having a hash), they lack the information necessary to either directly correlate you with any actual ballot or to brute-force what the ballot must have been to generate the hash. And since anyone can look to see if a name's been ticked or not off the voter's lists, that's no more information than they'd have at present.
That's one of a number of possible solutions to the veracity problem. Because there are many solutions to veracity, not all of which are compatible with the many solutions to other parts of the puzzle, it's not useful to focus on that one solution. What you ideally want to do is to start with the bits for which there are provably very few solutions because then you minimize the risk of producing flaws elsewhere by having to leave out parts.
Which is why there needs to be some method of making it provable that a vote is legitimate without violating the anonymity. In other words, the voter and the vote should be non-repudiatable in isolation but no combination of voter with vote should also be non-repudiatable. That's tough, in fact it's the single-toughest problem in the whole e-voting system, which is why I consider it to be the problem that needs to be solved first with all other components built around that solution. Everything else is trivial and therefore you can produce a near-infinite number of useless solutions by solving those bits first.
The e-voting system I've suggested a few times (anonymous generation of private/public key pairs issued to those who have ID) would make the showing of ID superfluous. You can vote with an invalid encryption key if you like, but there's bugger all the voting computer can do with it as it can't decrypt it. This also avoids the objections (which are valid) by individuals who have complained excessive ID requirements make voting impossible (in violation of the Federal laws on voting, not to mention the 15th Amendment which prohibits disenfranchisement of those who are legit, regardless of reason or excuse, no matter how good on the surface it may be).
The PET 3032 (back in 1978), free-running and unsynchronised, was capable of 30-seconds-per-year accuracy on a decent, clean power supply. That was, admittedly, about the absolute limit, but you could do it. A modern computer runs around 4 billion times as many cycles per second. More if you supercool then overclock it. A modern computer also has up to 16 cores per node and fairly typical clusters can have 64 nodes.
As for analog watches, the high-end mechanical watches you can buy off-the-shelf have a drift of around 1 second per day (30 times better than your estimate and 3 times better than any computer is capable of doing if the power supply will induce 3 seconds a day error). For free-running digital devices, a typical Casio quartz digital watch is around six nines accuracy (0.1 seconds drift a day), no synchronization required. Which means you can actually buy a cheap wristwatch that's 30x more accurate on timing than the best home computer you can get.
Sorry if I find the incompetence of hardware engineers a little hard to accept, I just prefer standards that, y'know, improve over time, not regress. 3 seconds a day drift is what vintage Swiss watches could do. I prefer modern technology to do better than the stuff that Huygens could do, not merely equal it.
20 minutes over the course of a year is a lot when you can make clocks accurate to within a second over 3.7 billion years. It's little only in relative terms for "typical usage". Hard real-time means as close to absolutely linear CPU time to wallclock time as you can achieve, which in turns means all locks and synchronizations must be absolutely predictable in advance and absolutely uniform in time. These kinds of systems often have nanosecond-accurate internal clocks for a reason, not for amusement. Add in random drifts of 20 minutes a year and what you have is a pile of spare parts 'cos it's useless for what it's intended for.
Of course, people generally don't run TeV-scale nuclear accelerators or hot fusion experiments in their home. Not that they would want to, but what they want doesn't matter. What they have to work with is incapable of doing the job.
And that, ultimately, is the biggest problem. You, as a geek/nerd/whatever CANNOT do anything to a higher standard than what the parts supplied will allow. How bright you are, how gifted you are, how free you imagine yourself to be - none of that matters. You can do NOTHING with greater accuracy than the sum of all cumulative errors will permit. Which is no big deal when the errors are insignificant. And, yes, there's a hell of a lot you can do on a home computer where 3 seconds a day drift is way, way too much.
Why? Timing isn't an issue. The drift in phase due to the thermal expansion and contraction of the materials carrying the power is a bit of a nuicense, but using better-grade materials (making behaviour more predictable and more controllable) would solve some of that and substations are quite capable of handling the marginal extra complexity of preventing errors from accumulating.
The added complexity is needed anyway as virtually every major blackout in history (including all the ones in recent times) have been due to crappy power routing, even crappier signalling of faults and absolutely pathetic to the power of crappy management of what signals are sent. A decent communications infrastructure, together with competent error handling and proper fault-tolerence, is absolutely essential if we're to avoid having the grid toast itself the next time a branch falls or a solar storm hits.
But if you're going to have that kind of oomph anyway, with all that it would take to make sure the complexity is not itself a weakness in the system, is it seriously too much to ask to add in the necessary analogue hardware to lock the phase at 60 Hz with zero deviation within any sane or rational level of measurement? Hell, if it weren't for the fact that two top analogue engineers have just died (one of a stroke, the other from a car accident), it would be a cakewalk to make it zero deviation within assorted insane levels of measurement. As it is, it's merely difficult enough to be interesting.
This was the basis of one of Fred Hoyle's many novels based on the science of the time. (He was an astrophysicist and used fiction as a means of exploring the implications of the science.) In short, his theory was that the supernova would cause such severe global change that it would essentially end civilization and borderline-end humanity.
Edwards Air Force Base is not Area 51. It's not even in the same State.
Second, do you know exactly how much the black budget is? Ok, do you want to guess the cost of the detector arrays involved, the number of false alarms or maintenance bills? The bill for that kind of level of security is probably in the order of a few mil per sqaure yard per year. I'm sure the Feds would be more than happy for you to pay them to install the system and run it. Oh, you DON'T want to pay for it? Well, an aiirwall is pretty cheap, fits into the taxes you're willing to stump up perfectly.
Naturally. It's far cheaper to lower everyone's standards of living than to raise them Oh, and chocolate rations are going up tomorrow to half of what they were yesterday.
The requirements for carrying identification/immigration paperwork are exactly the same as the federal laws.
Even if true (which it isn't), the criminal acts of one is not a justification of the criminal acts of another. And, yes, I regard authority figures going round demanding identification papers (something the US used to sneer at Russia over doing, I might add) as a criminal act. It is amusing, if ironic, that it is those who demand smaller government and greater freedom from interference that are also demanding that the government expands in order to do this....or did you expect these illegal immigrants to form neat lines at the nearest police station to hand themselves in? Perhaps, whilst the police are busy checking the papers, the thieves, muggers and drive-by shooters would like to take an ill-deserved lunch break. The added taxation would pay for the extra paperwork involved.
Slashdot Mirror
Exactly. There's many a license that is generally accepted (even by the OSI) as Open Source that doesn't permit free-as-in-freedom usage. Parallel to that, academia has long considered the contents of the articles to be "open" (ie: there's no restriction on knowledge, there's no restriction on copying, etc) but not "free" (ie: you can't plagarize - ie: produce a fork of - a research paper). If there's a definition of "open source", then the academic one is the one that is the most universal.
Exactly, which is why I specified "limited open source" to differentiate it from the more generic concept.
Well, not really, since Thrust SSC (their previous car) did 715 mph and therefore was already mach 1. They later discovered that the severe damage to the underbelly would have meant that even a few mph faster would have destroyed the car (and driver) utterly.
Well, having lived in South Carolina for a few years, I can say that the Civil War may not be a shooting war but as a Cold War it most certainly isn't over. Some of the rhetoric I was hearing was downright scary. Regardless, open ballots do lend themselves much more to partisanship and the kind of popularism Plato warns against in The Republic, which is why the open ballots in Congress are often abused by powerful lobby groups to cripple dissent and pluralism. Those voting are so terrified of the lobbyists and wannabe-warlords that actual negotiation and amicable settlements for the good of both sides has become almost impossible.
In isolation, you're correct, anonymity is a non-issue. But maintaining the veracity of the data without compromising the anonymity is a good deal harder. It's not unusual to find some sort of fraud (ballot-box stuffing, the mysterious vanishing-acts of ballot boxes, dubious counting, etc). It is extremely hard to find ways to prove that all legitimately-cast ballots are counted and that no illegitimate ballots are included, except by weakening the anonymity (since ballots cast without any anonymity can very easily be shown to either be legitimate or fraudulant).
What I wanted to do was establish if there was a way to maximize both, with the compromises necessary to achieve this coming entirely from the parts of the system that have no real impact on the outcome or its security, and to determine just what those maxima were given that you can't compromise the rest to infinity, it will fall over at some point.
Because it requires only the most basic competence at thinking to understand that net cost for a society is the gross cost minus the gross benefit. I'm sorry to hear you fail on the basic competence test. Your brainless delusions are what the rest of us call false economies.
(Infrastructure will last a century or more and the US has a population in excess of a quarter billion. This renders the instantaneous cost to any specific individual of absolutely no consequence, even if you assume that the instantaneous cost would indeed be that great to begin with. People building national power grids are probably not short of a dime or two.)
If you want it in terms simple enough for anyone with more than an amoeba for a brain to comprehend, the law of diminishing returns demands that putting more in will eventually get you less out, but most systems end up being S-curves, so there's actually a time when putting more in will get you greatly more out - and therefore putting less in will equally get you much less out.
The first four of those tend to vote Republican. Ask the Gipper.
Think of it as a SQL statement. If you start with the smallest table and join onto that, both you as a developer/tester and the computer will have the least work to do.
Ok, the smallest solution-space would seem to be to make each ballot unlinkable to a voter and yet be able to prove that the mapping of ballots to votes is a perfect 1:1, that all voters were authorized and that the ballots counted were the ones presented.
This is small because you have veracity of every set and every relationship at the same time as you have repudiatability of any connection between any vote and any voter. There simply aren't many ways you can meet those two heavily-conflicting constraints.
Here are the options I've been able to come up with. They all rely on voting machines not counting votes but merely doing a bit of trivial local processing with all the actual counting being done somewhere else. There is therefore an actual electronic ballot from start to finish, rather than a tally that could come from anything, and since a complete set of actual ballots exists, they can be recounted or inspected at any later date as is the case with a paper ballot (preserving the elements of veracity within the current solution). They also all provide methods by which third-party observers (including the general public) could be provided with sufficient information to monitor the voting for fraud at some level without violating the anonymity of the votes or the privacy of the voters.
(Indeed, it is assumed in all of these methods that you want limited-capacity third-party observers so that if holes were to be found they could not be exploited without detection, where the limited nature of the observing isn't itself compromising the security.)
1. Anonymous public/private encryption key pairs.
A voter casts an encrypted ballot in which the key they possess is useless to anyone wishing to find out what the vote was, but where there is one and only one key that can decrypt that ballot and produce a valid record. This requires that you have two machines - the one generating the key pair and the one doing the decryption, where both are tamper-proof, the link is unidirectional and the link is also tamper-proof. The one generating the key pair provides the human users with the encrypting key part only, which forms a part of the voter registration card used in the act of voting. The human users never see the decrypting key, which is passed solely to the decrypting system. With no linkage between the keys, even if you could snoop in on the communications you could never positively link any given encrypting key with any given decrypting key.
(The link therefore doesn't have to be snoop-proof, though that's obviously preferred, it only has to be tamper-proof so that it is never possible for an outsider to inject a false key pair into the system.)
If you place your vote directly onto the "voter registration card" (think more of what a single-function PDA would look like, given all the functions you can cram onto a smartphone today, than a simple piece of card with a name on it) and that card produces the encrypted ballot, the encryption key is never exposed and therefore cannot be duplicated and used by someone else.
If the decryption system destroys all keys that have already been used and ignores all ballots for which no decryption key exists on the system, no encryption key can ever be used for more than one vote and no vote for which the key generator did not generate the key pair will ever be counted.
You now have a system where one card is unique to one voter is unique to one ballot is unique to one recorded vote, provably so, without violating the confidentiality of any of the sequence.
This system has a limit on scalability. The decrypting machine is essentially using the decrypt key list that it has as a variant of a rainbow table. It must, therefore, try !(number of voters) possible keys in the worst possible case in order to be able to find the right key. (Since we know in advance th
The number of people who use a Casio watch for hard real-time computing is probably also small. My comment can be reduced to this: The best solution will ALWAYS allow for the greatest flexibility for the greatest number (within, as I've noted elsewhere, practical limits). Solutions that are constraining or inferior to what can be practically achieved for no reason whatsoever beyond the fact that the solver thinks that they can get away with it are not solutions worth having.
We can design a national grid system that has a drift far, far less than that described at no siginificant extra cost to what we need to make the national grid robust anyway. That reduced drift increases the value of the grid system and increases the scope of what will work. If it's done particularly well *cough*, then the cost of cleaning up the drift will not only be insignificant overall but will also be lower than the costs you would need to spend on individualized solutions for individual geeks, nerds, inventors and analogue freaks in order to achieve the same level of benefit.
As far as I'm concerned, I have no objection to spending less and getting more, even if it means someone, somewhere, has to be smarter than going for the cheapest, dirtiest hack job the market can stomach without puking its guts out.
People who sync their watches to the power grid are in for a shocking experience.
My point is not that people sync watches with power grids but that watches, clocks and computing needs already exist that are far more accurate than that and need to be, and that therefore it is irrational and naive to claim a 3 sec/day drift is acceptable, or that ANY given amount of drift is acceptable, because what is acceptable is defined by the problem and not some absolute.
The argument that drift/noise/aliens on the power lines is acceptable because it doesn't interfere in way X with device Y is likewise irrational and naive for the same reason. You can only say it is acceptable in some specific problem-space where the net result is within tolerable bounds (and no affect at all is a perfectly valid net result and a perfectly valid result to test against bounds). Any problem outside that problem-space is automatically undefined by that reasoning.
My argument reduces to this: NO individual, interested only in some specific problem-space, has the right to tell another individual that their problem-space is unworthy of solving or being solvable. Artificial, stupid and ultimately unnecessary limits should be avoided, not volunteered-for. Yes, the law of diminishing returns means there's a limit to what a practical solution can do, that's why I qualify my statement to those limits that are artificial AND stupid AND ultimately unnecessary, for if a solution is all three then it cannot possibly be the limit of what is practical.
You are correct that voters should never be attachable to a vote, but the prior poster is also correct that it is essential that it be provable that the votes counted were the ones cast and that all legitimate votes cast were counted. A sufficiently powerful cryptographic hash (perhaps with sufficient salt from the myriads of identification documents everyone has on file) might work. You have a hash, you can look up to see if the hash is listed amongst the votes counted, but all anyone else could do would establish that you voted (by merit of having a hash), they lack the information necessary to either directly correlate you with any actual ballot or to brute-force what the ballot must have been to generate the hash. And since anyone can look to see if a name's been ticked or not off the voter's lists, that's no more information than they'd have at present.
That's one of a number of possible solutions to the veracity problem. Because there are many solutions to veracity, not all of which are compatible with the many solutions to other parts of the puzzle, it's not useful to focus on that one solution. What you ideally want to do is to start with the bits for which there are provably very few solutions because then you minimize the risk of producing flaws elsewhere by having to leave out parts.
Which is why there needs to be some method of making it provable that a vote is legitimate without violating the anonymity. In other words, the voter and the vote should be non-repudiatable in isolation but no combination of voter with vote should also be non-repudiatable. That's tough, in fact it's the single-toughest problem in the whole e-voting system, which is why I consider it to be the problem that needs to be solved first with all other components built around that solution. Everything else is trivial and therefore you can produce a near-infinite number of useless solutions by solving those bits first.
The e-voting system I've suggested a few times (anonymous generation of private/public key pairs issued to those who have ID) would make the showing of ID superfluous. You can vote with an invalid encryption key if you like, but there's bugger all the voting computer can do with it as it can't decrypt it. This also avoids the objections (which are valid) by individuals who have complained excessive ID requirements make voting impossible (in violation of the Federal laws on voting, not to mention the 15th Amendment which prohibits disenfranchisement of those who are legit, regardless of reason or excuse, no matter how good on the surface it may be).
The PET 3032 (back in 1978), free-running and unsynchronised, was capable of 30-seconds-per-year accuracy on a decent, clean power supply. That was, admittedly, about the absolute limit, but you could do it. A modern computer runs around 4 billion times as many cycles per second. More if you supercool then overclock it. A modern computer also has up to 16 cores per node and fairly typical clusters can have 64 nodes.
As for analog watches, the high-end mechanical watches you can buy off-the-shelf have a drift of around 1 second per day (30 times better than your estimate and 3 times better than any computer is capable of doing if the power supply will induce 3 seconds a day error). For free-running digital devices, a typical Casio quartz digital watch is around six nines accuracy (0.1 seconds drift a day), no synchronization required. Which means you can actually buy a cheap wristwatch that's 30x more accurate on timing than the best home computer you can get.
Sorry if I find the incompetence of hardware engineers a little hard to accept, I just prefer standards that, y'know, improve over time, not regress. 3 seconds a day drift is what vintage Swiss watches could do. I prefer modern technology to do better than the stuff that Huygens could do, not merely equal it.
20 minutes over the course of a year is a lot when you can make clocks accurate to within a second over 3.7 billion years. It's little only in relative terms for "typical usage". Hard real-time means as close to absolutely linear CPU time to wallclock time as you can achieve, which in turns means all locks and synchronizations must be absolutely predictable in advance and absolutely uniform in time. These kinds of systems often have nanosecond-accurate internal clocks for a reason, not for amusement. Add in random drifts of 20 minutes a year and what you have is a pile of spare parts 'cos it's useless for what it's intended for.
Of course, people generally don't run TeV-scale nuclear accelerators or hot fusion experiments in their home. Not that they would want to, but what they want doesn't matter. What they have to work with is incapable of doing the job.
And that, ultimately, is the biggest problem. You, as a geek/nerd/whatever CANNOT do anything to a higher standard than what the parts supplied will allow. How bright you are, how gifted you are, how free you imagine yourself to be - none of that matters. You can do NOTHING with greater accuracy than the sum of all cumulative errors will permit. Which is no big deal when the errors are insignificant. And, yes, there's a hell of a lot you can do on a home computer where 3 seconds a day drift is way, way too much.
Why? Timing isn't an issue. The drift in phase due to the thermal expansion and contraction of the materials carrying the power is a bit of a nuicense, but using better-grade materials (making behaviour more predictable and more controllable) would solve some of that and substations are quite capable of handling the marginal extra complexity of preventing errors from accumulating.
The added complexity is needed anyway as virtually every major blackout in history (including all the ones in recent times) have been due to crappy power routing, even crappier signalling of faults and absolutely pathetic to the power of crappy management of what signals are sent. A decent communications infrastructure, together with competent error handling and proper fault-tolerence, is absolutely essential if we're to avoid having the grid toast itself the next time a branch falls or a solar storm hits.
But if you're going to have that kind of oomph anyway, with all that it would take to make sure the complexity is not itself a weakness in the system, is it seriously too much to ask to add in the necessary analogue hardware to lock the phase at 60 Hz with zero deviation within any sane or rational level of measurement? Hell, if it weren't for the fact that two top analogue engineers have just died (one of a stroke, the other from a car accident), it would be a cakewalk to make it zero deviation within assorted insane levels of measurement. As it is, it's merely difficult enough to be interesting.
Yeah, but the AC has an extra-laggy connection.
Well, yes! That should be obvious.
This was the basis of one of Fred Hoyle's many novels based on the science of the time. (He was an astrophysicist and used fiction as a means of exploring the implications of the science.) In short, his theory was that the supernova would cause such severe global change that it would essentially end civilization and borderline-end humanity.
Edwards Air Force Base is not Area 51. It's not even in the same State.
Second, do you know exactly how much the black budget is? Ok, do you want to guess the cost of the detector arrays involved, the number of false alarms or maintenance bills? The bill for that kind of level of security is probably in the order of a few mil per sqaure yard per year. I'm sure the Feds would be more than happy for you to pay them to install the system and run it. Oh, you DON'T want to pay for it? Well, an aiirwall is pretty cheap, fits into the taxes you're willing to stump up perfectly.
When this spreads to New Mexico, will Area 51 be exempt?
Naturally. It's far cheaper to lower everyone's standards of living than to raise them Oh, and chocolate rations are going up tomorrow to half of what they were yesterday.
The requirements for carrying identification/immigration paperwork are exactly the same as the federal laws.
Even if true (which it isn't), the criminal acts of one is not a justification of the criminal acts of another. And, yes, I regard authority figures going round demanding identification papers (something the US used to sneer at Russia over doing, I might add) as a criminal act. It is amusing, if ironic, that it is those who demand smaller government and greater freedom from interference that are also demanding that the government expands in order to do this. ...or did you expect these illegal immigrants to form neat lines at the nearest police station to hand themselves in? Perhaps, whilst the police are busy checking the papers, the thieves, muggers and drive-by shooters would like to take an ill-deserved lunch break. The added taxation would pay for the extra paperwork involved.