Most countries have their own boundaries and privacy expectations, very few of which match the general "ideal" American view. Streetview is only possible if culture is ignored, which can only serve to reinforce conservatism and isolationism, as the instinctive reaction to boundary violations is to repel the offender.
That would be my understanding as well. I believe early attempts at producing a vaccine tried to use de-activated HIV but it was actually able to reactivate itself, but you'd have to dig through the early work to get the low-down on that. The most recent efforts were abandoned the other week, with researchers being quoted as saying they had to go back to square one. NOT the simple problem as presented in Neuromancer! Despite claims to the contrary, there are no known examples of survivors and no known examples of anyone being immune. The claimed cases of marginal resistance to contracting the virus are, as far as I know, just claims and rumours. If there was any basis, at least that anyone could make use of, researchers would at least be that far along. The fact that they are so depressed suggests any such resistance is, in fact, futile.
The spread of the disease, the date of origin in humans, and its mutation rate have all recently been revised, suggesting that even the most basic science on the subject is proving extremely tough - far tougher than expected.
My thoughts on the subject aren't much different than they were when I first heard about this killer virus. Virus RNA is a molecule. There are properties of molecules (such as unique absorption lines) that make them possible to detect and attack, without having to care about what the molecule actually is or what the context is. The frequency used by your microwave oven is independent of where the hydrogen-oxygen bond is, and nothing else in the microwave oven will absorb on that specific line. Heat is subsequently transferred by convection, conduction and radiation, but the only points initially targeted are those hydrogen-oxygen bonds.
In the case of X-Ray fluorescence, you use the same technique, only at a much higher energy. It is certainly reversible, so if you shine the right frequency of X-Rays onto something, you get electrons, and if you shine electrons at the right energy, you get X-Rays. But, again, only from the isotope (this time) that corresponds to that very specific energy. Nothing else will interact with either.
In principle, this means that mutations don't matter so long as you can obtain the virus from the infected person. Since you will get all the current mutations in that person, it is utterly irrelevant as to whether there's one or a million. That just determines how many lines you need to emit over. Since the virus is fragile, the risks to the rest of the body from secondary radiation or some coincidence in absorption lines must be relatively small. Certainly smaller than certain death.
Maybe that particular approach is useless, and it very probably is, but at least it's no more useless than all the other methods people have tried, and at least tries to get past the problems of past failed approaches by tackling the problem from a different direction entirely.
I do, however, feel absolutely convinced that if a biochemical or biological solution could be found, it would have been. These guys are not idiots, they're not short on research funds, and they're certainly not short on material. Ergo, that solution domain is likely NOT where any real solutions lie. The answers will necessarily be in physics (work on the basis that we're concerned with a molecule) or mathematics (if prions can refold proteins, and if proteins are what prevent HIV from being identified, then refold the protective protein envelope until it's either recognized or will not function as intended). There are no other fields that are likely to be the least bit of interest.
Optimization is about minimizing some set of N functions out of some superset of N + M functions. If you assume that N=1 and that the function has exactly one minimum (eg: it is linear with a lower limit, or quadratic), then yes, there is one best solution. In practice, programs are not that simple. If you want to solve a more complex system - optimizing multiple parameters that vary non-linearly, which is the normal case - then you will have some number of local minima and one or more global minima.
(A local minimum is the lowest point within a region, but may not be the lowest point overall. Automatic optimizers - or indeed any automatic herustic - will tend to get stuck in local minima unless well programmed, as it will "appear" a good solution because of a lack of understanding of the problem space. Understanding, of course, is impossible until or unless a machine passes the Turing Test.)
In a few, rare, cases, no optimizer could ever solve the problem because there are no minima to find. This would happen if the relationship between the different characteristics you wish to optimize is a chaotic system which, when graphically mapped, would produce a fractal. Every point in a true fractal can be expanded to produce more information, and because the system is sensitive to initial conditions, that information cannot be inferred from any surrounding information. It is thus impossible to ever know if you are converging or diverging from a solutions, and impossible to know when such a solution is reached, since there is always an infinite amount of points that could potentially be better.
I shall leave it as an exercise to Slashdotters who are also on OpenCores to develop a microprocessor instruction set that is (a) Turing-Complete, but (b) possesses the nature that no non-trivial program can ever be optimized for the above reason. (ie: the limits of any function describing a characteristic are always Strange Attractors, not minima or maxima.)
If you buy a PC that uses a Linux BIOS but has XP installed, is that considered an XP sale, a Linux sale, or both? If both, then any of those Asus motherboards with a Linux-based BIOS should be added to the Linux column as well.
Well, only prophetic in that the Government, in conjunction with Gerry Anderson and the BBC Wardrobe Department, had an experimental time-traveling mind-reading ray.
Cobol was never a good language, and computer scientists fought hard to move the industry away from it because the probability of writing defective code was extremely high. It is readable, but really only in the same sense Intercal is. There are many, many hundreds of programming languages out there with far more programmers in each than Cobol has ever had. Use what makes real sense, not what some medieval history teacher thinks ancient civilizations may have liked. Hell, ADA has more merit - and more support for modern capabilities. If people are really serious about Cobol, I'm going to write an Algol65 plugin for Firefox - see how people like Algolscript pages. Nyah!
If a "legacy" language is used, the smartest choices would be Cilk or Occam, both of which make multicore and multinode processing trivial, and also make secure, robust programming practical. They SHOULD be on any sane IT manager's radar by now - the insane ones will go with COBOL and the notorious software unreliability that plagued the planet as a result.
Those wanting to go with a less drastic update really aught to be looking (by now) at OpenMP (which supports many languages) or Unified Parallel C (which at least lets you leverage existing C skills). Not sure where D is on parallelisms, but it is a cleaner OO language than most (C++ and Java especially take note) and might be helpful just for that.
If you absolutely insist on using existing ancient code, on the grounds that it has had longer to have bugs worked out, FORTRAN is the way to go. Businesses never pushed COBOL very hard, because it's so damn fragile. If you've ever had a trillion dollar water bill, that was a degenerate COBOL application. FORTRAN, by being used by high-end facilities where a single error could fry the entire budget for several years, tended to be used by people with considerably greater skill. So much so that it is still the number one programming language for mathematically-oriented programs, even though C maths libraries are as good or maybe better. The skill that is possible in FORTRAN is such that it is still the language of choice in many specialist sectors.
(Note: FORTRAN and COBOL are correctly given in upper case as they were invented before lower-case letters. It is often assumed that Admiral Grace Hopper was in the US Navy, when in fact the continental land-masses had not yet divided at the time she developed COBOL.)
Tin is a Government plot to leave people's minds exposed. True faraday cages use copper, at least 0.5" thick. Since the Chinese are in on this, and they're below us, it also has to be a full-body suit. I also recommend charging it with a 400,000 volt van de graaf generator to repel the purple ants from the planet Zog. To maintain the charge, remember to wear platform rubber-sole shoes.
The thing that cracks me up is that the one service I've not yet seen mentioned on Slashdot that is affected is exactly the one a geek might have figured on first - the practice of VPN tunneling over DNS servers. (See Freshmeat, as always, for details.) The attack obviously means such VPN tunnels can be spliced into. This means anything that can be reached by such tunnels, even if the endpoints concerned cannot be remotely accessed by any other means, are essentially wide open.
Now, I don't personally know of anyone who uses such tunneling software, but that's not the point. This is a GEEK site! Geeky but irrelevant vulnerabilities should rank higher than mundane, boring, obvious ones that most geeks should not care about anyway. (When I started running my own MUSH servers - I had 7 going at one point - I didn't trust external DNS servers to be safe, reliable or up-to-date, so simply zone dumped all the regulars onto my own DNS and ignored the outside DNS tree entirely. If anyone had problems, I re-transferred the zone from IP address, never name, and always from the authoritative source, never secondaries. These days, that could constitute breach of copyright or an act of terrier-ism, so I've stopped running MUSHes and MUDs.)
Reed-Solomon is designed only for randomly-distributed errors. For errors that are in a large, contiguous block (such as a sector) you need Turbo Codes. So, whilst you are correct that the solution isn't that useful for the bulk of real-world conditions (ie: lost drives), it really isn't useful for the condition it is supposed to fix either (lost sectors). As others have noted, most drives already employ Reed-Solomon to fix random bit errors, so employing it a second time would seem to just hog space with error-correcting codes that you don't need and can't effectively use. I'm not convinced employing Turbo Codes would be that useful either, but at least they'd fix errors not already being fixed.
That does depend on the nature of the encryption. A one-time pad is unbreakable, completely and utterly, provided the key is not directly available to the attacker. This is not theoretical, it is absolute. This is because a OTP will decrypt equally to EVERY possible message of the known length or shorter, with no possibility of knowing which of the apparent solutions is valid. Technically, they all are.
OTPs are a bugbear because you have to physically supply the decryption key by some method, but if that method was truly secure, then it would be equally possible to send the message that way. Workarounds include sending excessively large keys, then at message time, use a synchronization method to provide a random offset into the key. This is the method used by the British military's "C-Corp" (communications group). Makes random intercepts much much harder, as the key alone wouldn't be sufficient. More recent attempts have included using random radio sources from space as part of the key, as a short message using conventional encryption giving the coordinates is much much harder to crack. This is much harder than it sounds - radio telescopes need careful designing and placement to avoid terrestrial interference. That's not something you can implement just anywhere.
What of conventional encryption? Well, you obviously want to get as close to the characteristics of a one-time pad as possible - that is to say, it's less important as to whether an attacker finds the right key as it is that the attacker should not realize they have done so.
One popular method (though mathematically difficult to prove) is to use extremely heavy lossless compression first. The result will appear random. If you have chosen the compression algorithm well, tests for whether it is random or not should fall below whatever the accepted threshold is, so that it is rejected as a valid decrypt.
THEN you want to apply a strong encryption algorithm. If it is a block algorithm, the blocking mode should be non-trivial, so as to make it harder to attack potential vulnerabilities in the algorithm. There are many extremely good modes out there now. EAX, CCM and GCM are NIST-approved. OCB is also very reasonable. Use of AES as an encryption mode was not examined by NIST, but the proposal is on their site and it does sound interesting. A good mode without a good cipher is not terribly useful, and there are several ciphers listed on the block cipher lounge that do not have known weaknesses.
Yes, all of this is no better than the encryption key (or the seed + algorithm used to generate a key). A weak key is useless, as is a strong key if generated by a seed that can be deduced easily (a very common social engineering attack). However, there is really nothing to stop a competent user from ensuring that no decryption system will extract the original file in any meaningful timeframe, even if one of the methods listed above is broken. The more competent the user, the more of those methods have to have vulnerabilities.
The ideal is to make the total cost per message broken exceed the total value of the messages that have been broken. Then, you have added social engineering as an additional level of protection. If the probable cost is too high, nobody will bother, even though there is a random chance they'd hit the right solution immediately. It's similar to Douglas Adams' "Somebody Else's Problem Field".
Re:no encryption that YOU didn't write is safe
on
Is Hushmail Still Safe?
·
· Score: 3, Interesting
Upgrade the EFF's DES cracker to modern processors or GPU cores (whichever would be better at cracking DES), and decryption times of a few minutes would be realistically achievable. Depending on how efficient their code was (eg: could it fit entirely into L2, with data?), there may be room for improvement there. Add in superior cooling and overclocking techniques, you can probably get another 10-20% speedup. So if you really wanted, you could probably crack DES in under a minute, using off-the-shelf components. Triple DES is many orders of magnitude harder, I know of no machines currently out there that could make a serious dent in it. You'd need to find a weakness caused by how the DES algorithms interacted to mount a serious challenge using today's technology.
Boy, that's a surprise. Nova is a Latin word meaning "new", and both Spanish and English are Latin derivatives. (French has an interesting history in that it was considered degenerate Latin until the Revolution, at which point it became the language that had always been spoken and was unchanging and unchangeable. Go figure.)
Many urban legends are so patently stupid one must wonder if there isn't a team of anthropologists out there doing a hundred year study, seeing how such legends are picked up and transferred through and between communities. (If there isn't such a study, there damn well should be. Grant money for this can be sent to my regular PO Box.)
Nothing wrong with bragging, and Slashdot is known to cause superpower-generating mutations, which is why CowboyNeil does so well in the polls all the time.
Teleporters are crude. Since matter is just energy with an information matrix overlaid in which the physical location is part of that information, altering the matrix should alter the position of the matter without the need for a teleport system with its inherent problems of information bandwidth and Heisenberg uncertainty. Of course, that's not going to happen next year (or even another hundred years). Direct manipulation of the information that binds energy to form matter is unlikely to be possible for another 500 - 1000 years. Add another 50 - 60 before it becomes possible to use that ability to transport macroscale objects, such as people, safely and reliably.
Let's test my claims versus the examples you gave, see if my prediction is correct.
Cuba - current administration founded by violent revolution - maintained by violence and intimidation - results and proportion match theory
North Korea - founded by violent civil war - maintained by violence and intimidation - results and proportion match theory
Laos - current administration founded by violent civil war - maintained by force and suppression - results match theory
East Germany - founded by violent occupation of a society riddled with violence and xenophobia - maintained brutally by force and intimidation - results and proportion match theory
Soviet Union - founded by a combination of violent revolution and violent conquest - maintained by savage violence and intimidation - results and proportion match theory
Remember, my theory says NOTHING about success, wealth, power, influence or even freedoms. It states that a system will typically be maintained by the means by which it is created. You can have a violently maintained nation that still has enormous freedoms. In the most trivial sense, the fact that US cops carry guns is all about maintaining law and order through the threat of them shooting you. If that wasn't an effective threat, why would they bother?
You can also have FUD and freedom. Classic example - the raising of the national alert status for the Democratic Convention in 2004. Wonderful example of deliberate scare tactics, but it didn't alter anyone's freedom to attend. Margret Thatcher routinely used scare tactics to frighten people away from voting Labour, superb FUD work, but she never stopped anyone supporting them or voting for them.
My first claim is that you cannot have a nation that was founded through violence that uses neither FUD nor violence to survive - such methods guarantee real and imagined grievances will make it unsafe for such a nation to ever renounce such methods. Any nation that tries will inevitably get itself replaced. You have shown me no counter-example. All your examples verify this claim.
My second claim is that you cannot have a nation that was founded through peaceful means that uses either FUD or violence to survive - should it try, the backlash will always exceed its capacity to deal with both the rebellion and whatever caused it to adopt such tactics in the first place. There are rather fewer examples of such societies, but they have existed (Skara Brae is a good example, surviving 1,500 until finally being beaten by the environment) and they do exist (Iceland is considered the most peaceful in the world, has been a genuine democracy since 930AD and I don't recall it getting mentioned here for Big Brother tactics).
I make no other claims, although since you brought up America's wealth, I would point out Iceland is one of the wealthiest nations in the world. This would indicate to me that any nation is capable of wealth, that wealth is wholly independent of the means to sustain the government.
(Please note: I dislike Iceland's stance on many issues, but I respect them for being honest in their views. I wouldn't want to live there, and I do choose to live in the US freely even though I regard the current regime as a bunch of mindless thugs and don't expect any future regime to improve on that.)
You might want to look at what was highlighted - the PROVIDER is under investigation, not the tribe. It is the provider that is corrupt, the tribe (very likely) has no more technical knowledge or business acumen than any other non-technical non-corporate organization. ie: not much. This looks like a typical case of a business finding people who lack the necessary skills to evaluate a contract and decided to rip them off as much as possible before getting caught. Hell, I've worked for multinationals that are incapable of evaluating contracts and got themselves screwed over. If you can't expect Fortune 500 companies to bother reading what is written, just because of a fancy powerpoint presentation, can you seriously expect a community get-together to do better?
That's a very simplistic viewpoint and runs counter to historical experience and the majority of accepted non-partisan political theory. It is a viewpoint that mostly goes along with governments that use Fear, Uncertainty and Doubt (FUD) as weapons of political terrorism. The United States does an exceptional job at creating FUD. Indeed, every government created by violence or fear has perpetuated itself through violence and fear, whereas governments formed by peaceful consent almost never use either violence or fear, they survive by consent perfectly well.
(I'm not going to argue over whether or not the war of independence was necessary or not, to me that is irrelevant. What matters is that the war is still going on in the minds of those who run the country, and all is "fair" in love and war - even when the war is a delusionary one.)
The "easiest" solution would be to run a bunch of fibres to some "neutral" point on each block. Although this uses multiple cables, with one cable per end-point in the junction box, it's the same distribution mechanism that cable currently uses. (You see cable junction boxes on some telephone poles, but also as small green pedestals in front of houses and as junction boxes on the sides of apartment buildings.) The "last mile" becomes the "last few feet", with the cable relatively easy (and therefore relatively cheap) to reach and replace.
If you wanted to do municipal/metropolitan broadband, you'd have 32 fibres run to each block, then a 256-way multicast-capable, MPLS-capable router linking four blocks together. (MPLS, or some other virtual circuit protocol, would then uniquely tag a user's stream, so it can be identified further along.) This would be linked to a switch, in the case of larger cities, which would link up a fairly large set of these 4-blocks into a well-defined subset of the city. You'd then have a set of top level multicast-capable MPLS-capable routers that linked the layer below it onto the public Internet, possibly through multiple gateways. Residents would then "buy" Internet access from the providers as always, but this would only require adjusting a QoS table entry in one top-level router that identified how much bandwidth a given virtual circuit had on the public Internet and which gateway that connection would use.
For intra-city connections - say, IMing a friend in the next building - you would only go over the metronet, and your connection could sensibly be whatever speed the local fibre could handle - call it a gigabit per second - provided the upstream networks weren't saturated, as you're working over shared pipes some of the way. Saturation can be avoided by placing routers and switches in parallel. You could load-balance between them, or you could have them working wholly in parallel and have very high-speed switches linking the independent metronets together into a collective metronet. In either case, it makes no difference which router a packet comes in on or goes out on, even if the routers are not on the same "tree" per-se.
If you don't have limited funds, then saturation is inevitable at some point. To minimize the overall impact, routers should be enabled with CBQ or HFSC, such that each virtual circuit has a guaranteed bandwidth (something it can always reach, no matter how busy the network) and a hard maximum bandwidth of whatever the local few meters connection can support, where the guaranteed bandwidth is either an equal fraction of the network at that segment or the hard maximum, whichever is less.
Could this be done? Yes. It's not anti-competitive, as ISPs still end up selling bandwidth to customers the way they have always done. The metronet doesn't replace the ISPs, it replaces the need for excessive physical wiring and it allows ISPs that provide broadband to do so without buying/maintaining quite so many expensive DSL modems, so it cuts the ISP's costs.
Is such a model in use? Yes. It's how natural gas and electricity are sold already. It's how DSL works, for the most part, as DSL companies all share the same phone lines. The difference is the line supplier, not the principle.
Hatfill had many years to assemble and prepare a case before it got to court, a fair amount of money for (a) a decent lawyer, and (b) living expenses when under repressive conditions. Since that time, however, US citizens have been arrested over terrorism within the US, denied Constitutional rights, quite possibly tortured, and then convicted on unrelated charges despite a fairly decent chance of insanity. In this case, the researcher may well have been certifiable due to possible exposure to neurotoxins. If that is indeed the case, the Government had no business pursuing the death penalty even if the guy was guilty. If being exposed to brain-mushing chemicals and having your cerebral cortex swiss-cheezed is not grounds for insanity, what the hell is?
However, we then come to the other side of the case. People exposed to neurotoxins on that kind of level (eg: those exposed to mercury, in the hatting industry) are generally incapable of functioning much beyond survival. Planning and executing attacks so stealthily and so meticulously as to not even be a serious suspect for so many years - well, I have a hard time believing that anyone affected to that degree would be physically or mentally able to do something like that.
Part three of this ramble is the whole presumption of innocence. We, as a nation, have long agreed on the principles that an accusation is not the same as a conviction, and that even a conviction can be in error. A conviction is only reasonable grounds for assuming guilt, it is not a guarantee of it. (This is why the term "reasonable doubt" is used in trials, not "absolute certainty".) However, this guy had not even been officially charged, let alone convicted. As such, there is absolutely no grounds for claiming he is guilty beyond reasonable doubt. Even if both sides present what evidence they have to a court of some kind, without the accused being able to assist in the defense, it's hard to see how you can even reach the giddy heights of "balance of probability", the lowest legal standard that exists outside of Guantanamo.
Let's say there is some sort of independent investigation and the guy is found innocent. You think this Government or the next would be willing to pay the next of kin a few million dollars for reckless endangerment (I consider the way this was handled to have been exactly that), libel, slander and damage to his professional reputation? Yeah, right. If he was cleared tomorrow, the Government is more likely to send the bill for their press conference to the estate than a check.
Slashdot Mirror
Most countries have their own boundaries and privacy expectations, very few of which match the general "ideal" American view. Streetview is only possible if culture is ignored, which can only serve to reinforce conservatism and isolationism, as the instinctive reaction to boundary violations is to repel the offender.
That would be my understanding as well. I believe early attempts at producing a vaccine tried to use de-activated HIV but it was actually able to reactivate itself, but you'd have to dig through the early work to get the low-down on that. The most recent efforts were abandoned the other week, with researchers being quoted as saying they had to go back to square one. NOT the simple problem as presented in Neuromancer! Despite claims to the contrary, there are no known examples of survivors and no known examples of anyone being immune. The claimed cases of marginal resistance to contracting the virus are, as far as I know, just claims and rumours. If there was any basis, at least that anyone could make use of, researchers would at least be that far along. The fact that they are so depressed suggests any such resistance is, in fact, futile.
The spread of the disease, the date of origin in humans, and its mutation rate have all recently been revised, suggesting that even the most basic science on the subject is proving extremely tough - far tougher than expected.
My thoughts on the subject aren't much different than they were when I first heard about this killer virus. Virus RNA is a molecule. There are properties of molecules (such as unique absorption lines) that make them possible to detect and attack, without having to care about what the molecule actually is or what the context is. The frequency used by your microwave oven is independent of where the hydrogen-oxygen bond is, and nothing else in the microwave oven will absorb on that specific line. Heat is subsequently transferred by convection, conduction and radiation, but the only points initially targeted are those hydrogen-oxygen bonds.
In the case of X-Ray fluorescence, you use the same technique, only at a much higher energy. It is certainly reversible, so if you shine the right frequency of X-Rays onto something, you get electrons, and if you shine electrons at the right energy, you get X-Rays. But, again, only from the isotope (this time) that corresponds to that very specific energy. Nothing else will interact with either.
In principle, this means that mutations don't matter so long as you can obtain the virus from the infected person. Since you will get all the current mutations in that person, it is utterly irrelevant as to whether there's one or a million. That just determines how many lines you need to emit over. Since the virus is fragile, the risks to the rest of the body from secondary radiation or some coincidence in absorption lines must be relatively small. Certainly smaller than certain death.
Maybe that particular approach is useless, and it very probably is, but at least it's no more useless than all the other methods people have tried, and at least tries to get past the problems of past failed approaches by tackling the problem from a different direction entirely.
I do, however, feel absolutely convinced that if a biochemical or biological solution could be found, it would have been. These guys are not idiots, they're not short on research funds, and they're certainly not short on material. Ergo, that solution domain is likely NOT where any real solutions lie. The answers will necessarily be in physics (work on the basis that we're concerned with a molecule) or mathematics (if prions can refold proteins, and if proteins are what prevent HIV from being identified, then refold the protective protein envelope until it's either recognized or will not function as intended). There are no other fields that are likely to be the least bit of interest.
Optimization is about minimizing some set of N functions out of some superset of N + M functions. If you assume that N=1 and that the function has exactly one minimum (eg: it is linear with a lower limit, or quadratic), then yes, there is one best solution. In practice, programs are not that simple. If you want to solve a more complex system - optimizing multiple parameters that vary non-linearly, which is the normal case - then you will have some number of local minima and one or more global minima.
(A local minimum is the lowest point within a region, but may not be the lowest point overall. Automatic optimizers - or indeed any automatic herustic - will tend to get stuck in local minima unless well programmed, as it will "appear" a good solution because of a lack of understanding of the problem space. Understanding, of course, is impossible until or unless a machine passes the Turing Test.)
In a few, rare, cases, no optimizer could ever solve the problem because there are no minima to find. This would happen if the relationship between the different characteristics you wish to optimize is a chaotic system which, when graphically mapped, would produce a fractal. Every point in a true fractal can be expanded to produce more information, and because the system is sensitive to initial conditions, that information cannot be inferred from any surrounding information. It is thus impossible to ever know if you are converging or diverging from a solutions, and impossible to know when such a solution is reached, since there is always an infinite amount of points that could potentially be better.
I shall leave it as an exercise to Slashdotters who are also on OpenCores to develop a microprocessor instruction set that is (a) Turing-Complete, but (b) possesses the nature that no non-trivial program can ever be optimized for the above reason. (ie: the limits of any function describing a characteristic are always Strange Attractors, not minima or maxima.)
If you buy a PC that uses a Linux BIOS but has XP installed, is that considered an XP sale, a Linux sale, or both? If both, then any of those Asus motherboards with a Linux-based BIOS should be added to the Linux column as well.
If the new ones run Vista, it would explain why nobody in the country has any money any more.
Well, only prophetic in that the Government, in conjunction with Gerry Anderson and the BBC Wardrobe Department, had an experimental time-traveling mind-reading ray.
Cobol was never a good language, and computer scientists fought hard to move the industry away from it because the probability of writing defective code was extremely high. It is readable, but really only in the same sense Intercal is. There are many, many hundreds of programming languages out there with far more programmers in each than Cobol has ever had. Use what makes real sense, not what some medieval history teacher thinks ancient civilizations may have liked. Hell, ADA has more merit - and more support for modern capabilities. If people are really serious about Cobol, I'm going to write an Algol65 plugin for Firefox - see how people like Algolscript pages. Nyah!
If a "legacy" language is used, the smartest choices would be Cilk or Occam, both of which make multicore and multinode processing trivial, and also make secure, robust programming practical. They SHOULD be on any sane IT manager's radar by now - the insane ones will go with COBOL and the notorious software unreliability that plagued the planet as a result.
Those wanting to go with a less drastic update really aught to be looking (by now) at OpenMP (which supports many languages) or Unified Parallel C (which at least lets you leverage existing C skills). Not sure where D is on parallelisms, but it is a cleaner OO language than most (C++ and Java especially take note) and might be helpful just for that.
If you absolutely insist on using existing ancient code, on the grounds that it has had longer to have bugs worked out, FORTRAN is the way to go. Businesses never pushed COBOL very hard, because it's so damn fragile. If you've ever had a trillion dollar water bill, that was a degenerate COBOL application. FORTRAN, by being used by high-end facilities where a single error could fry the entire budget for several years, tended to be used by people with considerably greater skill. So much so that it is still the number one programming language for mathematically-oriented programs, even though C maths libraries are as good or maybe better. The skill that is possible in FORTRAN is such that it is still the language of choice in many specialist sectors.
(Note: FORTRAN and COBOL are correctly given in upper case as they were invented before lower-case letters. It is often assumed that Admiral Grace Hopper was in the US Navy, when in fact the continental land-masses had not yet divided at the time she developed COBOL.)
Press releases are aways written in really weird English, so I've found a converter to turn it into something more readable.
Tin is a Government plot to leave people's minds exposed. True faraday cages use copper, at least 0.5" thick. Since the Chinese are in on this, and they're below us, it also has to be a full-body suit. I also recommend charging it with a 400,000 volt van de graaf generator to repel the purple ants from the planet Zog. To maintain the charge, remember to wear platform rubber-sole shoes.
The thing that cracks me up is that the one service I've not yet seen mentioned on Slashdot that is affected is exactly the one a geek might have figured on first - the practice of VPN tunneling over DNS servers. (See Freshmeat, as always, for details.) The attack obviously means such VPN tunnels can be spliced into. This means anything that can be reached by such tunnels, even if the endpoints concerned cannot be remotely accessed by any other means, are essentially wide open.
Now, I don't personally know of anyone who uses such tunneling software, but that's not the point. This is a GEEK site! Geeky but irrelevant vulnerabilities should rank higher than mundane, boring, obvious ones that most geeks should not care about anyway. (When I started running my own MUSH servers - I had 7 going at one point - I didn't trust external DNS servers to be safe, reliable or up-to-date, so simply zone dumped all the regulars onto my own DNS and ignored the outside DNS tree entirely. If anyone had problems, I re-transferred the zone from IP address, never name, and always from the authoritative source, never secondaries. These days, that could constitute breach of copyright or an act of terrier-ism, so I've stopped running MUSHes and MUDs.)
Suspended Animation offers better recovery possibilities, especially now the latest suspend-to-disk has got into the kernel.
Reed-Solomon is designed only for randomly-distributed errors. For errors that are in a large, contiguous block (such as a sector) you need Turbo Codes. So, whilst you are correct that the solution isn't that useful for the bulk of real-world conditions (ie: lost drives), it really isn't useful for the condition it is supposed to fix either (lost sectors). As others have noted, most drives already employ Reed-Solomon to fix random bit errors, so employing it a second time would seem to just hog space with error-correcting codes that you don't need and can't effectively use. I'm not convinced employing Turbo Codes would be that useful either, but at least they'd fix errors not already being fixed.
That does depend on the nature of the encryption. A one-time pad is unbreakable, completely and utterly, provided the key is not directly available to the attacker. This is not theoretical, it is absolute. This is because a OTP will decrypt equally to EVERY possible message of the known length or shorter, with no possibility of knowing which of the apparent solutions is valid. Technically, they all are.
OTPs are a bugbear because you have to physically supply the decryption key by some method, but if that method was truly secure, then it would be equally possible to send the message that way. Workarounds include sending excessively large keys, then at message time, use a synchronization method to provide a random offset into the key. This is the method used by the British military's "C-Corp" (communications group). Makes random intercepts much much harder, as the key alone wouldn't be sufficient. More recent attempts have included using random radio sources from space as part of the key, as a short message using conventional encryption giving the coordinates is much much harder to crack. This is much harder than it sounds - radio telescopes need careful designing and placement to avoid terrestrial interference. That's not something you can implement just anywhere.
What of conventional encryption? Well, you obviously want to get as close to the characteristics of a one-time pad as possible - that is to say, it's less important as to whether an attacker finds the right key as it is that the attacker should not realize they have done so.
One popular method (though mathematically difficult to prove) is to use extremely heavy lossless compression first. The result will appear random. If you have chosen the compression algorithm well, tests for whether it is random or not should fall below whatever the accepted threshold is, so that it is rejected as a valid decrypt.
THEN you want to apply a strong encryption algorithm. If it is a block algorithm, the blocking mode should be non-trivial, so as to make it harder to attack potential vulnerabilities in the algorithm. There are many extremely good modes out there now. EAX, CCM and GCM are NIST-approved. OCB is also very reasonable. Use of AES as an encryption mode was not examined by NIST, but the proposal is on their site and it does sound interesting. A good mode without a good cipher is not terribly useful, and there are several ciphers listed on the block cipher lounge that do not have known weaknesses.
Yes, all of this is no better than the encryption key (or the seed + algorithm used to generate a key). A weak key is useless, as is a strong key if generated by a seed that can be deduced easily (a very common social engineering attack). However, there is really nothing to stop a competent user from ensuring that no decryption system will extract the original file in any meaningful timeframe, even if one of the methods listed above is broken. The more competent the user, the more of those methods have to have vulnerabilities.
The ideal is to make the total cost per message broken exceed the total value of the messages that have been broken. Then, you have added social engineering as an additional level of protection. If the probable cost is too high, nobody will bother, even though there is a random chance they'd hit the right solution immediately. It's similar to Douglas Adams' "Somebody Else's Problem Field".
Upgrade the EFF's DES cracker to modern processors or GPU cores (whichever would be better at cracking DES), and decryption times of a few minutes would be realistically achievable. Depending on how efficient their code was (eg: could it fit entirely into L2, with data?), there may be room for improvement there. Add in superior cooling and overclocking techniques, you can probably get another 10-20% speedup. So if you really wanted, you could probably crack DES in under a minute, using off-the-shelf components. Triple DES is many orders of magnitude harder, I know of no machines currently out there that could make a serious dent in it. You'd need to find a weakness caused by how the DES algorithms interacted to mount a serious challenge using today's technology.
Boy, that's a surprise. Nova is a Latin word meaning "new", and both Spanish and English are Latin derivatives. (French has an interesting history in that it was considered degenerate Latin until the Revolution, at which point it became the language that had always been spoken and was unchanging and unchangeable. Go figure.)
Many urban legends are so patently stupid one must wonder if there isn't a team of anthropologists out there doing a hundred year study, seeing how such legends are picked up and transferred through and between communities. (If there isn't such a study, there damn well should be. Grant money for this can be sent to my regular PO Box.)
Nothing wrong with bragging, and Slashdot is known to cause superpower-generating mutations, which is why CowboyNeil does so well in the polls all the time.
Well, there WAS one single-engined competitor, but alas it was in the UK, not the US.
Teleporters are crude. Since matter is just energy with an information matrix overlaid in which the physical location is part of that information, altering the matrix should alter the position of the matter without the need for a teleport system with its inherent problems of information bandwidth and Heisenberg uncertainty. Of course, that's not going to happen next year (or even another hundred years). Direct manipulation of the information that binds energy to form matter is unlikely to be possible for another 500 - 1000 years. Add another 50 - 60 before it becomes possible to use that ability to transport macroscale objects, such as people, safely and reliably.
Remember, my theory says NOTHING about success, wealth, power, influence or even freedoms. It states that a system will typically be maintained by the means by which it is created. You can have a violently maintained nation that still has enormous freedoms. In the most trivial sense, the fact that US cops carry guns is all about maintaining law and order through the threat of them shooting you. If that wasn't an effective threat, why would they bother?
You can also have FUD and freedom. Classic example - the raising of the national alert status for the Democratic Convention in 2004. Wonderful example of deliberate scare tactics, but it didn't alter anyone's freedom to attend. Margret Thatcher routinely used scare tactics to frighten people away from voting Labour, superb FUD work, but she never stopped anyone supporting them or voting for them.
My first claim is that you cannot have a nation that was founded through violence that uses neither FUD nor violence to survive - such methods guarantee real and imagined grievances will make it unsafe for such a nation to ever renounce such methods. Any nation that tries will inevitably get itself replaced. You have shown me no counter-example. All your examples verify this claim.
My second claim is that you cannot have a nation that was founded through peaceful means that uses either FUD or violence to survive - should it try, the backlash will always exceed its capacity to deal with both the rebellion and whatever caused it to adopt such tactics in the first place. There are rather fewer examples of such societies, but they have existed (Skara Brae is a good example, surviving 1,500 until finally being beaten by the environment) and they do exist (Iceland is considered the most peaceful in the world, has been a genuine democracy since 930AD and I don't recall it getting mentioned here for Big Brother tactics).
I make no other claims, although since you brought up America's wealth, I would point out Iceland is one of the wealthiest nations in the world. This would indicate to me that any nation is capable of wealth, that wealth is wholly independent of the means to sustain the government.
(Please note: I dislike Iceland's stance on many issues, but I respect them for being honest in their views. I wouldn't want to live there, and I do choose to live in the US freely even though I regard the current regime as a bunch of mindless thugs and don't expect any future regime to improve on that.)
You might want to look at what was highlighted - the PROVIDER is under investigation, not the tribe. It is the provider that is corrupt, the tribe (very likely) has no more technical knowledge or business acumen than any other non-technical non-corporate organization. ie: not much. This looks like a typical case of a business finding people who lack the necessary skills to evaluate a contract and decided to rip them off as much as possible before getting caught. Hell, I've worked for multinationals that are incapable of evaluating contracts and got themselves screwed over. If you can't expect Fortune 500 companies to bother reading what is written, just because of a fancy powerpoint presentation, can you seriously expect a community get-together to do better?
That's a very simplistic viewpoint and runs counter to historical experience and the majority of accepted non-partisan political theory. It is a viewpoint that mostly goes along with governments that use Fear, Uncertainty and Doubt (FUD) as weapons of political terrorism. The United States does an exceptional job at creating FUD. Indeed, every government created by violence or fear has perpetuated itself through violence and fear, whereas governments formed by peaceful consent almost never use either violence or fear, they survive by consent perfectly well.
(I'm not going to argue over whether or not the war of independence was necessary or not, to me that is irrelevant. What matters is that the war is still going on in the minds of those who run the country, and all is "fair" in love and war - even when the war is a delusionary one.)
The "easiest" solution would be to run a bunch of fibres to some "neutral" point on each block. Although this uses multiple cables, with one cable per end-point in the junction box, it's the same distribution mechanism that cable currently uses. (You see cable junction boxes on some telephone poles, but also as small green pedestals in front of houses and as junction boxes on the sides of apartment buildings.) The "last mile" becomes the "last few feet", with the cable relatively easy (and therefore relatively cheap) to reach and replace.
If you wanted to do municipal/metropolitan broadband, you'd have 32 fibres run to each block, then a 256-way multicast-capable, MPLS-capable router linking four blocks together. (MPLS, or some other virtual circuit protocol, would then uniquely tag a user's stream, so it can be identified further along.) This would be linked to a switch, in the case of larger cities, which would link up a fairly large set of these 4-blocks into a well-defined subset of the city. You'd then have a set of top level multicast-capable MPLS-capable routers that linked the layer below it onto the public Internet, possibly through multiple gateways. Residents would then "buy" Internet access from the providers as always, but this would only require adjusting a QoS table entry in one top-level router that identified how much bandwidth a given virtual circuit had on the public Internet and which gateway that connection would use.
For intra-city connections - say, IMing a friend in the next building - you would only go over the metronet, and your connection could sensibly be whatever speed the local fibre could handle - call it a gigabit per second - provided the upstream networks weren't saturated, as you're working over shared pipes some of the way. Saturation can be avoided by placing routers and switches in parallel. You could load-balance between them, or you could have them working wholly in parallel and have very high-speed switches linking the independent metronets together into a collective metronet. In either case, it makes no difference which router a packet comes in on or goes out on, even if the routers are not on the same "tree" per-se.
If you don't have limited funds, then saturation is inevitable at some point. To minimize the overall impact, routers should be enabled with CBQ or HFSC, such that each virtual circuit has a guaranteed bandwidth (something it can always reach, no matter how busy the network) and a hard maximum bandwidth of whatever the local few meters connection can support, where the guaranteed bandwidth is either an equal fraction of the network at that segment or the hard maximum, whichever is less.
Could this be done? Yes. It's not anti-competitive, as ISPs still end up selling bandwidth to customers the way they have always done. The metronet doesn't replace the ISPs, it replaces the need for excessive physical wiring and it allows ISPs that provide broadband to do so without buying/maintaining quite so many expensive DSL modems, so it cuts the ISP's costs.
Is such a model in use? Yes. It's how natural gas and electricity are sold already. It's how DSL works, for the most part, as DSL companies all share the same phone lines. The difference is the line supplier, not the principle.
What, he files in the 6th book? What page?
Hatfill had many years to assemble and prepare a case before it got to court, a fair amount of money for (a) a decent lawyer, and (b) living expenses when under repressive conditions. Since that time, however, US citizens have been arrested over terrorism within the US, denied Constitutional rights, quite possibly tortured, and then convicted on unrelated charges despite a fairly decent chance of insanity. In this case, the researcher may well have been certifiable due to possible exposure to neurotoxins. If that is indeed the case, the Government had no business pursuing the death penalty even if the guy was guilty. If being exposed to brain-mushing chemicals and having your cerebral cortex swiss-cheezed is not grounds for insanity, what the hell is?
However, we then come to the other side of the case. People exposed to neurotoxins on that kind of level (eg: those exposed to mercury, in the hatting industry) are generally incapable of functioning much beyond survival. Planning and executing attacks so stealthily and so meticulously as to not even be a serious suspect for so many years - well, I have a hard time believing that anyone affected to that degree would be physically or mentally able to do something like that.
Part three of this ramble is the whole presumption of innocence. We, as a nation, have long agreed on the principles that an accusation is not the same as a conviction, and that even a conviction can be in error. A conviction is only reasonable grounds for assuming guilt, it is not a guarantee of it. (This is why the term "reasonable doubt" is used in trials, not "absolute certainty".) However, this guy had not even been officially charged, let alone convicted. As such, there is absolutely no grounds for claiming he is guilty beyond reasonable doubt. Even if both sides present what evidence they have to a court of some kind, without the accused being able to assist in the defense, it's hard to see how you can even reach the giddy heights of "balance of probability", the lowest legal standard that exists outside of Guantanamo.
Let's say there is some sort of independent investigation and the guy is found innocent. You think this Government or the next would be willing to pay the next of kin a few million dollars for reckless endangerment (I consider the way this was handled to have been exactly that), libel, slander and damage to his professional reputation? Yeah, right. If he was cleared tomorrow, the Government is more likely to send the bill for their press conference to the estate than a check.