Well, gcc is likely to keep being the world's de facto C compiler (though even this was mainly because of the egcs fork way back when).
Actually, I doubt that is true. At this point, the commercial UNIX vendors and the BSDs seem to be putting their weight behind Clang/LLVM/LLDB, in large part due to GCC going GPLv3. In addition to being a cleaner architecture that's easier to enhance than GCC, it is also faster, and it often produces much better code as well. The GNU toolchain's days as the de facto standard are numbered, IMHO.
Back on topic, it occurs to be that large clusters with hundreds of cores start to inherently behave a lot more like NUMA and really need to be treated that way. Note that lots of modern OSes, including Linux, have supported NUMA in the past, so suggesting that it requires a completely rewritten OS is a preposterous assertion. That's not at all what this article is saying. What this article is saying is that tasks often are not easily divisible into tasks small enough to take advantage of multiple cores, and that managing processor affinity to ensure that threads working on the same data are run on the cores within the same physical die starts to become an unmanageable problem past a certain point.
In effect, what it is saying is that barring interconnect improvements, for many classes of problems, the performance penalty caused by multiple cores needing to access the same data exceeds the performance gain from adding additional cores at or around 48 cores. No OS change will help this, and in many cases, no software changes can help this, either. Most computing tasks are simply not massively parallelizable. This conclusion should be entirely expected by anybody who has ever tried to parallelize software to any real degree, but it's always good to see studies that bear out.
Put another way, once you exceed about 48 cores, the cores start to act more like clusters than cores. You start to see more and more accesses in which one CPU has to force data out of another CPU's cache. The nonuniformity of memory accesses starts to dominate the access times. Thus, past about that point (and probably much lower for most problems), adding more cores no longer improves performance. Even for massively parallelizable problems like video compression, once you exceed a certain number of nodes doing the work, the time spent assembling the final data actually exceeds the performance win achieved by adding additional processing nodes. This is completely straightforward, completely understood by real-world computer programmers, and shouldn't really be a surprise to anyone.
I'm not convinced an OS change can fix this, nor even an architectural change, though both can help to some degree by making parallelization easier (e.g. by providing APIs for supporting work units arranged in a dependency graph like GCD as an alternative to raw thread-based APIs). At some point, though, you're bounded by the number of distinct pieces that a problem can be divided into that don't depend on the output of any other piece, and once you hit that limit, adding additional computational units can only hinder performance, not help it. Your only real choices, then, are to find new and interesting ways to refactor the problem so that this is no longer the case, to change the structure of the input data to remove dependencies, to increase the speed of the individual CPU cores, or to turn the machines loose processing more than one problem at any given time to keep the remaining cores occupied.
Oh, yeah, and there's one other change that helps a lot: keep your read-only data in read-only pages, and write your code so that results go somewhere else. Read-only pages can be cached in every CPU without any real cache coherency overhead, at least in theory (I'm assuming that most modern CPUs do this), which means that input data sharing between CPUs doesn't matter. This design, combined with lockless work unit APIs, can make a huge difference in how many CPU
That's the thing, the automated lanes could safely do 120+ under normal driving conditions. Only a manual lane would need to have restrictive speed limits.
The sole reason for low speed limits is that humans suck at paying attention to driving conditions. A car could actively monitor its own tread depth and its slippage on a given road, monitor the level of slippage of other vehicles on the road ahead of it, and could use that plus the radius of the curve to calculate a maximum safe speed for every inch of the road at a level that only the best human race car drivers could pull off. And because cars would be able to communicate with each other, you would never have an accident caused by a car pulling out in front of another car. The cars would simply collaborate to time their lane changes and acceleration to maximize traffic flow while allowing the new vehicle to merge onto the road in a timely manner.
If all vehicles were automated, the accident rate would plunge because quite frankly, almost all accidents are either directly caused by driver error. Even most of the accidents that aren't caused by human error could generally be avoided if cars were capable of communicating mechanical failures to other cars on the road. Faster and safer.
Although i've yet to see a corporation want to join the military, and am not quite sure how that would even work.
I'm guessing it would go something like this: a company who wants to reduce headcount gets a huge grant from the government in exchange for the company enlisting and sending half their employees over to be cannon fodder.
And braking time is only a tiny portion of what controls whether you have an accident. Drunk drivers also swerve all over the road, then drive off the road and hit things. Most drunk driving accidents are not caused by traffic in front of them coming to a complete stop and them taking too long to respond. In fact, that occurs so rarely on most highways that it's almost completely irrelevant.
The fact is that reaction time only matters when something goes catastrophically wrong. Good drivers (read: sober) can generally take other precautions to avoid getting into situations where their reaction times will matter. Thus, if you are driving correctly (leaving proper distance between you and the car in front of you, etc.), driving while talking on the phone is no more dangerous than driving without talking on the phone.
Distractions are sometimes avoidable, sometimes not. If people don't know how to properly compensate for the level of distraction, distractions will continue to cause accidents, but this need not be the case. What we need to be doing is educating people on how to properly drive when distracted instead of just teaching them the laughably unrealistic and overly simplistic rule that they should avoid distractions. Distracted driver training should be a mandatory part of every driver ed curriculum. The basic rules are:
Be sure to tell the person on the other end of any phone call that you are in a car so that they know that you may have to suddenly stop talking and listening.
Do not allow distractions on any road that you are unfamiliar with. Take control as needed.
Keep double the normal stopping distance.
Slow down well in advance of traffic lights in case they change on you.
As distraction increases, your speed should decrease. Let the other drivers pass you. If they cut right in front of you, put on your brakes immediately to increase the space.
Watch for tricky situations. Cars passing you at a rapid speed mean that something could go very wrong very quickly up ahead. Same goes for brake lights in the distance.
When you notice a potentially tricky situation starting to form, say "hold on a minute" and drop the phone. Stop talking. Stop listening.
Do not attempt to call anyone back if a call drops on a difficult section of the road. It is better to wait than to wreck.
Do not attempt to answer questions that require significant thinking. Tell them you'll get back to them when you're not steering a four-wheeled bomb.
If you do all of these things religiously, they will more than make up for any deficit in attention caused by talking on the phone.
The bigger problem is human drivers. All these laws will become moot once cars can drive themselves, and we're already well on our way to that point. Within twenty years, these laws will be as quaint as laws regulating hanging up your boxer shorts on a clothesline on Sunday. Why bother passing laws to ban activity now when we're just going to have to fight to repeal those laws in single-digit years when they are no longer relevant?
Au contraire. You're mistakenly assuming that wiretaps are an end unto themselves. In fact, they are a means to an end, and those ends are often pretty horrible, up to and including executions. Since those executions would not have occurred without the wiretaps, the wiretaps are, in effect, about as egregious as you can get.
Put another way, a bolt fails on a tricycle because of poor manufacturing. The wheel falls off. A second bold fails on a bridge because of poor manufacturing. A school bus falls off. The first bolt manufacturing error was a minor infraction. The second, due to its eventual consequences, was a horrible act.
Nobody fully trusts any government. That's not really the point. There are governments that we trust more or less than others. I generally trust the U.S. government to mostly do something sane at least 75% of the time. I generally trust the Iranian government to do something sane at least 7.5% of the time. And therein was the point. It's not that we're a bunch of nutjobs who distrust the government and think that they're all out to get us and will abuse wiretapping authority frequently. It's that if the power exists, it will be abused, inevitably, by someone, against someone.
In the end, it doesn't matter whether the abuser is a rogue element in our government who isn't playing by the rules or a foreign government who doesn't have those rules in the first place. The result is the same. I just have a lot more faith that if it gets abused by some part of the U.S. government, there's at least some reasonable chance that they will eventually get caught and nailed to a wall over it.
This is the biggest reason why we fight against greater wiretap rules in the U.S. It's not that we don't trust our government, but rather that we can't trust all governments, and we're talking about world standards here. If we allow the U.S. government to put in rules that allow it to spy on Iranian citizens, due to the nature of the technology, we're also allowing Iran's government to spy on U.S. citizens. No matter how you look at it, it's pretty hard to argue that this is a good thing.
How many of those works are still making a significant profit for their creators after 14 years? 28? With the exception of novels, and possibly painting reproductions, a creative work has a shelf life of single-digit years. And even in those two cases, the average shelf life is pretty short; the ones that continue to turn a profit for decades are few and far between. Yet we deprive the public domain of these works that might otherwise see a resurgence in popularity under the false assumption that they might see a resurgence in popularity anyway. I think 28 years is way too long in the digital age, given the amount of content that is being created every day.
I think that copyright should be defined based on both time and profit. Once the work has brought in a certain reasonable baseline income, defined based on the nature of the work, and regularly adjusted for inflation, then the work should be protected for ten years, period, with no possibility for renewal. This would mean that lesser-known works would have the opportunity to take a while to build up in popularity and eventually pay back their creator for the time invested, but would also mean that a highly popular work would fall into the public domain about the time that its sales cease to be significant.
As long as there is transparency, and as long as the individuals get to choose whether their contributions get used for a particular purpose, that's fine. Groups should not have power in and of themselves, however. If I leave a PAC because they have drifted too far from their original purpose, I should have the right to take back my contribution. If I decide that my union is doing something stupid, I should be able to reduce their contribution by 1/200,000th. If I decide that the company I work for is making an inappropriate contribution with money that I helped create, I should be able to reduce that contribution to a degree commensurate with my contribution. See how hard that would be to handle, though? And because these groups cannot feasibly be made transparent, it is necessary to limit their ability to abuse their power.
Except that they can't. A corporation isn't a large group of individuals that decided to get together. It is an organization of a company that is typically owned by a small number of individuals (as few as one) for the purposes of protecting the assets of the proprietors if the company fails and to protect the proprietors from legal responsibility if the company does something in appropriate (usually). There is no good reason for a corporation to be treated as a person, as the sole purpose of a corporation is to prevent the individuals that make it up from having liability. Granting those people rights without liability attached is fundamentally unjust.
Corporate contributions to political campaigns, for example, means that someone in a position to dictate policy for a corporation is able to spend the money of other people without them really having any say in the matter. I mean sure, they could ostensibly sell their shares, but that just means that a different group of people are getting screwed. More to the point, that wealth was acquired because of the workers that make up the corporation, yet they have no real say whatsoever in how the money is spent except insofar as some of them may also be minority shareholders. In effect, this means that the voices of a few are amplified unfairly due solely to their being in charge of a corporation. In addition to these people casting one vote through their personal financial contributions, they are able to cast a second vote through corporate contributions, whereas the average citizen cannot realistically do the same. This promotes inequality in which the people with the most money and power are able to exert undue influence, thus increasing their money and power, in a sort of perverse feedback loop. This is exactly contrary to the founding principle of democracy---the notion that all are created equal abd should have equal rights under the law.
Unions are only slightly better. In principle, people have the right to refuse to join a union and can merely pay their "fair share" dues---the collective bargaining portion of the union dues without the political portion. In practice, however, the unions stand for many things, and there is no practical way for a union member to say that they will allow their dues to be spent on some, but not all of its political goals. As such, because it is a "take it or leave it" proposition and because contributions from a union represent substantially greater weight than individual contributions, members are unlikely to deny everything merely because they disagree with some of the union's positions. This again means that the decision-making process is taken out of the hands of its members and given to its elected leaders, again unfairly exaggerating the voice of a few on many issues.
I'd be fine with all the corporate and union contributions if I could say on an individual basis that their contribution must be reduced by 0.002% to account for me withholding my portion thereof. As long as this is not the case---as long as I don't have a vote on each individual issue---then corporations and unions do not accurately reflect the will of the people who comprise them, and as such, those contributions are a fundamental usurpation of power, denying us our rights as shareholders, union members, and workers in those corporations.
My point was to say your post appeared to have the argument we aren't winning this battle so lets allow it because it would be easier and cheaper at this point. One has to worry if such response were the case then at what point do we have to apply that with other laws.
The line I typically draw is that laws should cover harm to others and harm to society, and should go no further. Laws against burglary, murder, etc. are clear cases of preventing harm to others. Laws against using drugs that cause people to become dangerous addicts (meth, PCP, etc.) fall squarely on the good side of the line as well.
Laws about less harmful substances like pot, alcohol, and tobacco should be only as restrictive as is necessary to prevent the harm. For example, laws against smoking where others are forced to breathe the air, laws against driving drunk, laws against smoking on public beaches resulting from decades of litter problems by jerks who don't care, etc. fall squarely on the good side of the line because they are designed to limit behavior that causes harm to others.
Laws banning the substances altogether fall squarely on the other side of that line, as mere possession of any substance without using it cannot cause harm (unless you're talking about something radioactive or poisonous without proper containment) and these substances are not dangerous enough when used to warrant such restrictive laws. Even laws against the use of such substances within the privacy of your own home, assuming you don't have kids in your home, are unreasonably restrictive; the injury to others and to society can be better mitigated in other ways.
However, marijuana doesn't fit such a profile and has only gained I feel because so many people are now wanting it and combine some states are allowing "medical" marijuana.
Marijuana use was out of control long before medical marijuana came into play. Here are some stats on the popularity of marijuana. Notice that the numbers for young people were dropping until the early 90s, then started climbing again. The numbers for older people followed along a few years afterwards as those people got older and "aged up" into higher categories. Any effects of medical marijuana legislation (passed in 1996) should have started showing up after 1996, or at a minimum, should have started rising much faster, but in fact, the largest jump in youth use occurred before 1996. An effect cannot predate a cause, to the best of my knowledge. If anything, this suggests that perhaps the reverse is true---that laws to allow medical marijuana use stemmed from a resurgence in and growing acceptance of pot use.
More to the point, in people age 25-34, almost half have admitted using marijuana at some point in their lives. This has been true fairly consistently in every measurement beginning in the 1970s and continuing every 3 years up to the present. What's interesting about this is that people over 35 show lower lifetime pot use even after enough time has gone by for those people to "age up" into the next bracket, which basically suggests that people over 35 like to pretend that they never used marijuana because of social stigmas and/or because they have kids and don't want their kids to know that they smoked pot. Either way, it's a rather fascinating pile of statistics when you really start to think about it.:-)
I'd like to propose that we immediately start making plans for our resistance movement. A master key for an SSL/TLS session does not inherently mean the private key for the machine. After all, that's not the key used in the crypto, but merely used as part of the key exchange process.
I propose that we immediately start designing software features so that we can send the government the transaction keys using RSA/DSA for the initial connection. That way, every website can upload tens of thousands of keys per second to the government key escrow site and turn the whole thing into the farce that it is.
Another sad argument... If they are going to do it anyway then make it legal.
That is not my argument. My argument is that if a substance is of comparable harm to dozens of other substances that are all legal, AND if people are going to use that substance anyway, that prohibition is doomed to failure.
The singling out of marijuana as a Schedule I controlled substance comically exceeds any rational evaluation based on its long-term health effects, its addictiveness, or any other reasonable criteria. If marijuana is classified as schedule I, cigarettes should be even more highly restricted, as should alcohol, as both show substantially greater potential for clinical addiction and at least as much long term harm. Yet neither of these is a controlled substance. Hence, by any rational evaluation, marijuana should also not be a controlled substance, and any decision to make it one is entirely arbitrary, driven by political aims (and the lack of big companies backing it), rather than being driven by science and reason.
Your argument might hold up if I were arguing for legalization of PCP or meth, but I would never be foolish enough to make such an argument. Clearly, burglary causes significantly more harm than any legal activity that I'm aware of. Thus, comparing it to marijuana use is at best fallacious, and at worst, deliberately misleading, specious reasoning.
I know folks who are big on heroin etc now. They all stated they would never go past pot, but in truth it was the 'gateway' for them. While it sucks for the people that can indulge in a little pot from time to time overall I think making it illegal benefits wider society (although harshly punishing casual use does seem a bit extreme).
Actually, if marijuana truly is acting as a gateway drug, that's all the more reason to legalize it. I didn't bring up that point because the debate over the concept of "gateway drugs" is highly contentious at best.
All of your acquaintances who moved from marijuana to something else did so because they already knew a dealer who dealt other stuff, or at least knew people who did. If they were buying pot from legal dealers instead of on the black market, that relationship---that connection---would not exist, and thus those people would be much less likely to move on to harder drugs.
Except that it doesn't work. Random drug tests might stop people from having jobs, but it won't stop people from using drugs. There are far too many counterexamples to believe that enforcement against users will work.
Bear in mind that I'm not suggesting plea bargains just for end users. Work your way up the chain. Eventually, you'll have enough small-time dealers willing to rat on the next person up the chain to get a conviction, rinse, repeat until you've gotten everybody, at least on the U.S. side of the border. Repeat this often enough and make the chances of getting caught high enough, and people won't be nearly as willing to risk dealing just to make a quick buck.
I'm not arguing for legalizing all drugs. I'm arguing solely for marijuana because it seems to be lower on the "wrecks your body for life" scale than many legal substances. If young people are going to do it anyway, we should at least have a drug policy that sets consistent standards for what is and isn't illegal based on reasonable metrics of risk. As long as drugs that are relatively benign (I'm not saying marijuana is safe---smoking anything is inherently bad for your health---just that it's nowhere near as bad as meth or PCP) are illegal, young people think, "oh, the people who made those rules just don't understand," and they proceed to ignore the rules. Make the rules cover the hard stuff while going easy on the light stuff, and you'll be less likely to encourage drug use due to the ambivalence and rebelliousness of youth.
You're grossly oversimplifying things. A lot of factors have contributed:
Laws forbidding foreign ownership of property.
A government that does little to combat abject poverty.
Brain drain to the U.S. and other countries.
Bad people who prey upon the poor to be drug mules, growers, etc.
Ineffectual police enforcement in Mexico.
A U.S. drug policy that encourages black market trade rather than controlled trade.
Utter failure on the part of the U.S. government to combat abject poverty.
Bad people who prey upon the poor and offer them a better life through dealing drugs.
Ineffectual district attorneys who would rather "get tough on drug users" than offer plea bargains in exchange for ratting out their pushers (the original purpose of prohibiting use of these drugs).
Ineffectual police enforcement that similarly focuses on busting users instead of dealers.
There's plenty of blame to spread around on both sides of the fence. I do agree, though, that the best way to end drug violence is to create a legal marketplace for the least harmful and most common of those drugs. Prohibition never works if you're talking about products that people want to consume. You'd think the government would have learned this eighty years ago. The only way they got the U.S. back under control was by repealing prohibition. Sadly, the "morally superior" never learn. They just keep standing there in their ivory towers issuing edicts, repeating the same mistakes, and wondering why the side of the tower is burning.
That's okay, so long as the way it is applied makes sense based on the law, and so long as the law is written in such a way that the intent of the law is clear. Sadly, that latter part is quite often not the case in practice.
It's only overstressed because the people running the show are too busy grandstanding and politicking to actually get anything done. If Congress operated at even a tenth the efficiency of an equivalently sized body in private enterprise, they'd have plenty of time to get stuff done. Instead, they're too busy fighting with each other to do what we're paying them to do.
Put another way, Wi-Fi-enabled devices are predicted to cross the 200 million mark within 4 years. Multiply times even a ten megabyte file, and they'd be passing almost 2 petabytes of data per day. I don't think they've thought this through. That's about a tenth the traffic of Google or Steam. They're either completely s**t-for-brains stupid or they're trying to guarantee that this fails....
Slashdot Mirror
Actually, I doubt that is true. At this point, the commercial UNIX vendors and the BSDs seem to be putting their weight behind Clang/LLVM/LLDB, in large part due to GCC going GPLv3. In addition to being a cleaner architecture that's easier to enhance than GCC, it is also faster, and it often produces much better code as well. The GNU toolchain's days as the de facto standard are numbered, IMHO.
Back on topic, it occurs to be that large clusters with hundreds of cores start to inherently behave a lot more like NUMA and really need to be treated that way. Note that lots of modern OSes, including Linux, have supported NUMA in the past, so suggesting that it requires a completely rewritten OS is a preposterous assertion. That's not at all what this article is saying. What this article is saying is that tasks often are not easily divisible into tasks small enough to take advantage of multiple cores, and that managing processor affinity to ensure that threads working on the same data are run on the cores within the same physical die starts to become an unmanageable problem past a certain point.
In effect, what it is saying is that barring interconnect improvements, for many classes of problems, the performance penalty caused by multiple cores needing to access the same data exceeds the performance gain from adding additional cores at or around 48 cores. No OS change will help this, and in many cases, no software changes can help this, either. Most computing tasks are simply not massively parallelizable. This conclusion should be entirely expected by anybody who has ever tried to parallelize software to any real degree, but it's always good to see studies that bear out.
Put another way, once you exceed about 48 cores, the cores start to act more like clusters than cores. You start to see more and more accesses in which one CPU has to force data out of another CPU's cache. The nonuniformity of memory accesses starts to dominate the access times. Thus, past about that point (and probably much lower for most problems), adding more cores no longer improves performance. Even for massively parallelizable problems like video compression, once you exceed a certain number of nodes doing the work, the time spent assembling the final data actually exceeds the performance win achieved by adding additional processing nodes. This is completely straightforward, completely understood by real-world computer programmers, and shouldn't really be a surprise to anyone.
I'm not convinced an OS change can fix this, nor even an architectural change, though both can help to some degree by making parallelization easier (e.g. by providing APIs for supporting work units arranged in a dependency graph like GCD as an alternative to raw thread-based APIs). At some point, though, you're bounded by the number of distinct pieces that a problem can be divided into that don't depend on the output of any other piece, and once you hit that limit, adding additional computational units can only hinder performance, not help it. Your only real choices, then, are to find new and interesting ways to refactor the problem so that this is no longer the case, to change the structure of the input data to remove dependencies, to increase the speed of the individual CPU cores, or to turn the machines loose processing more than one problem at any given time to keep the remaining cores occupied.
Oh, yeah, and there's one other change that helps a lot: keep your read-only data in read-only pages, and write your code so that results go somewhere else. Read-only pages can be cached in every CPU without any real cache coherency overhead, at least in theory (I'm assuming that most modern CPUs do this), which means that input data sharing between CPUs doesn't matter. This design, combined with lockless work unit APIs, can make a huge difference in how many CPU
That's the thing, the automated lanes could safely do 120+ under normal driving conditions. Only a manual lane would need to have restrictive speed limits.
The sole reason for low speed limits is that humans suck at paying attention to driving conditions. A car could actively monitor its own tread depth and its slippage on a given road, monitor the level of slippage of other vehicles on the road ahead of it, and could use that plus the radius of the curve to calculate a maximum safe speed for every inch of the road at a level that only the best human race car drivers could pull off. And because cars would be able to communicate with each other, you would never have an accident caused by a car pulling out in front of another car. The cars would simply collaborate to time their lane changes and acceleration to maximize traffic flow while allowing the new vehicle to merge onto the road in a timely manner.
If all vehicles were automated, the accident rate would plunge because quite frankly, almost all accidents are either directly caused by driver error. Even most of the accidents that aren't caused by human error could generally be avoided if cars were capable of communicating mechanical failures to other cars on the road. Faster and safer.
I'm guessing it would go something like this: a company who wants to reduce headcount gets a huge grant from the government in exchange for the company enlisting and sending half their employees over to be cannon fodder.
And braking time is only a tiny portion of what controls whether you have an accident. Drunk drivers also swerve all over the road, then drive off the road and hit things. Most drunk driving accidents are not caused by traffic in front of them coming to a complete stop and them taking too long to respond. In fact, that occurs so rarely on most highways that it's almost completely irrelevant.
The fact is that reaction time only matters when something goes catastrophically wrong. Good drivers (read: sober) can generally take other precautions to avoid getting into situations where their reaction times will matter. Thus, if you are driving correctly (leaving proper distance between you and the car in front of you, etc.), driving while talking on the phone is no more dangerous than driving without talking on the phone.
Distractions are sometimes avoidable, sometimes not. If people don't know how to properly compensate for the level of distraction, distractions will continue to cause accidents, but this need not be the case. What we need to be doing is educating people on how to properly drive when distracted instead of just teaching them the laughably unrealistic and overly simplistic rule that they should avoid distractions. Distracted driver training should be a mandatory part of every driver ed curriculum. The basic rules are:
If you do all of these things religiously, they will more than make up for any deficit in attention caused by talking on the phone.
The bigger problem is human drivers. All these laws will become moot once cars can drive themselves, and we're already well on our way to that point. Within twenty years, these laws will be as quaint as laws regulating hanging up your boxer shorts on a clothesline on Sunday. Why bother passing laws to ban activity now when we're just going to have to fight to repeal those laws in single-digit years when they are no longer relevant?
That's for watching ChatRoulette.
It could be ironic if her name happens to be Chardonnay and you're trying to tell her that she isn't getting enough fly.... Or something....
Au contraire. You're mistakenly assuming that wiretaps are an end unto themselves. In fact, they are a means to an end, and those ends are often pretty horrible, up to and including executions. Since those executions would not have occurred without the wiretaps, the wiretaps are, in effect, about as egregious as you can get.
Put another way, a bolt fails on a tricycle because of poor manufacturing. The wheel falls off. A second bold fails on a bridge because of poor manufacturing. A school bus falls off. The first bolt manufacturing error was a minor infraction. The second, due to its eventual consequences, was a horrible act.
Nobody fully trusts any government. That's not really the point. There are governments that we trust more or less than others. I generally trust the U.S. government to mostly do something sane at least 75% of the time. I generally trust the Iranian government to do something sane at least 7.5% of the time. And therein was the point. It's not that we're a bunch of nutjobs who distrust the government and think that they're all out to get us and will abuse wiretapping authority frequently. It's that if the power exists, it will be abused, inevitably, by someone, against someone.
In the end, it doesn't matter whether the abuser is a rogue element in our government who isn't playing by the rules or a foreign government who doesn't have those rules in the first place. The result is the same. I just have a lot more faith that if it gets abused by some part of the U.S. government, there's at least some reasonable chance that they will eventually get caught and nailed to a wall over it.
This is the biggest reason why we fight against greater wiretap rules in the U.S. It's not that we don't trust our government, but rather that we can't trust all governments, and we're talking about world standards here. If we allow the U.S. government to put in rules that allow it to spy on Iranian citizens, due to the nature of the technology, we're also allowing Iran's government to spy on U.S. citizens. No matter how you look at it, it's pretty hard to argue that this is a good thing.
Hence the reason I changed the second sentence in that paragraph to "income" rather than "profit". The former is unambiguous.
No, more like:
Sandwich bag: $0.15
Bag overhead fee: $3.80 (the rest of the box)
Emergency courier fee: $15,000 (guy running a block to the 7-11)
How many of those works are still making a significant profit for their creators after 14 years? 28? With the exception of novels, and possibly painting reproductions, a creative work has a shelf life of single-digit years. And even in those two cases, the average shelf life is pretty short; the ones that continue to turn a profit for decades are few and far between. Yet we deprive the public domain of these works that might otherwise see a resurgence in popularity under the false assumption that they might see a resurgence in popularity anyway. I think 28 years is way too long in the digital age, given the amount of content that is being created every day.
I think that copyright should be defined based on both time and profit. Once the work has brought in a certain reasonable baseline income, defined based on the nature of the work, and regularly adjusted for inflation, then the work should be protected for ten years, period, with no possibility for renewal. This would mean that lesser-known works would have the opportunity to take a while to build up in popularity and eventually pay back their creator for the time invested, but would also mean that a highly popular work would fall into the public domain about the time that its sales cease to be significant.
As long as there is transparency, and as long as the individuals get to choose whether their contributions get used for a particular purpose, that's fine. Groups should not have power in and of themselves, however. If I leave a PAC because they have drifted too far from their original purpose, I should have the right to take back my contribution. If I decide that my union is doing something stupid, I should be able to reduce their contribution by 1/200,000th. If I decide that the company I work for is making an inappropriate contribution with money that I helped create, I should be able to reduce that contribution to a degree commensurate with my contribution. See how hard that would be to handle, though? And because these groups cannot feasibly be made transparent, it is necessary to limit their ability to abuse their power.
Except that they can't. A corporation isn't a large group of individuals that decided to get together. It is an organization of a company that is typically owned by a small number of individuals (as few as one) for the purposes of protecting the assets of the proprietors if the company fails and to protect the proprietors from legal responsibility if the company does something in appropriate (usually). There is no good reason for a corporation to be treated as a person, as the sole purpose of a corporation is to prevent the individuals that make it up from having liability. Granting those people rights without liability attached is fundamentally unjust.
Corporate contributions to political campaigns, for example, means that someone in a position to dictate policy for a corporation is able to spend the money of other people without them really having any say in the matter. I mean sure, they could ostensibly sell their shares, but that just means that a different group of people are getting screwed. More to the point, that wealth was acquired because of the workers that make up the corporation, yet they have no real say whatsoever in how the money is spent except insofar as some of them may also be minority shareholders. In effect, this means that the voices of a few are amplified unfairly due solely to their being in charge of a corporation. In addition to these people casting one vote through their personal financial contributions, they are able to cast a second vote through corporate contributions, whereas the average citizen cannot realistically do the same. This promotes inequality in which the people with the most money and power are able to exert undue influence, thus increasing their money and power, in a sort of perverse feedback loop. This is exactly contrary to the founding principle of democracy---the notion that all are created equal abd should have equal rights under the law.
Unions are only slightly better. In principle, people have the right to refuse to join a union and can merely pay their "fair share" dues---the collective bargaining portion of the union dues without the political portion. In practice, however, the unions stand for many things, and there is no practical way for a union member to say that they will allow their dues to be spent on some, but not all of its political goals. As such, because it is a "take it or leave it" proposition and because contributions from a union represent substantially greater weight than individual contributions, members are unlikely to deny everything merely because they disagree with some of the union's positions. This again means that the decision-making process is taken out of the hands of its members and given to its elected leaders, again unfairly exaggerating the voice of a few on many issues.
I'd be fine with all the corporate and union contributions if I could say on an individual basis that their contribution must be reduced by 0.002% to account for me withholding my portion thereof. As long as this is not the case---as long as I don't have a vote on each individual issue---then corporations and unions do not accurately reflect the will of the people who comprise them, and as such, those contributions are a fundamental usurpation of power, denying us our rights as shareholders, union members, and workers in those corporations.
The line I typically draw is that laws should cover harm to others and harm to society, and should go no further. Laws against burglary, murder, etc. are clear cases of preventing harm to others. Laws against using drugs that cause people to become dangerous addicts (meth, PCP, etc.) fall squarely on the good side of the line as well.
Laws about less harmful substances like pot, alcohol, and tobacco should be only as restrictive as is necessary to prevent the harm. For example, laws against smoking where others are forced to breathe the air, laws against driving drunk, laws against smoking on public beaches resulting from decades of litter problems by jerks who don't care, etc. fall squarely on the good side of the line because they are designed to limit behavior that causes harm to others.
Laws banning the substances altogether fall squarely on the other side of that line, as mere possession of any substance without using it cannot cause harm (unless you're talking about something radioactive or poisonous without proper containment) and these substances are not dangerous enough when used to warrant such restrictive laws. Even laws against the use of such substances within the privacy of your own home, assuming you don't have kids in your home, are unreasonably restrictive; the injury to others and to society can be better mitigated in other ways.
Marijuana use was out of control long before medical marijuana came into play. Here are some stats on the popularity of marijuana. Notice that the numbers for young people were dropping until the early 90s, then started climbing again. The numbers for older people followed along a few years afterwards as those people got older and "aged up" into higher categories. Any effects of medical marijuana legislation (passed in 1996) should have started showing up after 1996, or at a minimum, should have started rising much faster, but in fact, the largest jump in youth use occurred before 1996. An effect cannot predate a cause, to the best of my knowledge. If anything, this suggests that perhaps the reverse is true---that laws to allow medical marijuana use stemmed from a resurgence in and growing acceptance of pot use.
More to the point, in people age 25-34, almost half have admitted using marijuana at some point in their lives. This has been true fairly consistently in every measurement beginning in the 1970s and continuing every 3 years up to the present. What's interesting about this is that people over 35 show lower lifetime pot use even after enough time has gone by for those people to "age up" into the next bracket, which basically suggests that people over 35 like to pretend that they never used marijuana because of social stigmas and/or because they have kids and don't want their kids to know that they smoked pot. Either way, it's a rather fascinating pile of statistics when you really start to think about it. :-)
I'd like to propose that we immediately start making plans for our resistance movement. A master key for an SSL/TLS session does not inherently mean the private key for the machine. After all, that's not the key used in the crypto, but merely used as part of the key exchange process.
I propose that we immediately start designing software features so that we can send the government the transaction keys using RSA/DSA for the initial connection. That way, every website can upload tens of thousands of keys per second to the government key escrow site and turn the whole thing into the farce that it is.
That is not my argument. My argument is that if a substance is of comparable harm to dozens of other substances that are all legal, AND if people are going to use that substance anyway, that prohibition is doomed to failure.
The singling out of marijuana as a Schedule I controlled substance comically exceeds any rational evaluation based on its long-term health effects, its addictiveness, or any other reasonable criteria. If marijuana is classified as schedule I, cigarettes should be even more highly restricted, as should alcohol, as both show substantially greater potential for clinical addiction and at least as much long term harm. Yet neither of these is a controlled substance. Hence, by any rational evaluation, marijuana should also not be a controlled substance, and any decision to make it one is entirely arbitrary, driven by political aims (and the lack of big companies backing it), rather than being driven by science and reason.
Your argument might hold up if I were arguing for legalization of PCP or meth, but I would never be foolish enough to make such an argument. Clearly, burglary causes significantly more harm than any legal activity that I'm aware of. Thus, comparing it to marijuana use is at best fallacious, and at worst, deliberately misleading, specious reasoning.
Actually, if marijuana truly is acting as a gateway drug, that's all the more reason to legalize it. I didn't bring up that point because the debate over the concept of "gateway drugs" is highly contentious at best.
All of your acquaintances who moved from marijuana to something else did so because they already knew a dealer who dealt other stuff, or at least knew people who did. If they were buying pot from legal dealers instead of on the black market, that relationship---that connection---would not exist, and thus those people would be much less likely to move on to harder drugs.
Except that it doesn't work. Random drug tests might stop people from having jobs, but it won't stop people from using drugs. There are far too many counterexamples to believe that enforcement against users will work.
Bear in mind that I'm not suggesting plea bargains just for end users. Work your way up the chain. Eventually, you'll have enough small-time dealers willing to rat on the next person up the chain to get a conviction, rinse, repeat until you've gotten everybody, at least on the U.S. side of the border. Repeat this often enough and make the chances of getting caught high enough, and people won't be nearly as willing to risk dealing just to make a quick buck.
I'm not arguing for legalizing all drugs. I'm arguing solely for marijuana because it seems to be lower on the "wrecks your body for life" scale than many legal substances. If young people are going to do it anyway, we should at least have a drug policy that sets consistent standards for what is and isn't illegal based on reasonable metrics of risk. As long as drugs that are relatively benign (I'm not saying marijuana is safe---smoking anything is inherently bad for your health---just that it's nowhere near as bad as meth or PCP) are illegal, young people think, "oh, the people who made those rules just don't understand," and they proceed to ignore the rules. Make the rules cover the hard stuff while going easy on the light stuff, and you'll be less likely to encourage drug use due to the ambivalence and rebelliousness of youth.
You're grossly oversimplifying things. A lot of factors have contributed:
There's plenty of blame to spread around on both sides of the fence. I do agree, though, that the best way to end drug violence is to create a legal marketplace for the least harmful and most common of those drugs. Prohibition never works if you're talking about products that people want to consume. You'd think the government would have learned this eighty years ago. The only way they got the U.S. back under control was by repealing prohibition. Sadly, the "morally superior" never learn. They just keep standing there in their ivory towers issuing edicts, repeating the same mistakes, and wondering why the side of the tower is burning.
That's okay, so long as the way it is applied makes sense based on the law, and so long as the law is written in such a way that the intent of the law is clear. Sadly, that latter part is quite often not the case in practice.
It's only overstressed because the people running the show are too busy grandstanding and politicking to actually get anything done. If Congress operated at even a tenth the efficiency of an equivalently sized body in private enterprise, they'd have plenty of time to get stuff done. Instead, they're too busy fighting with each other to do what we're paying them to do.
Put another way, Wi-Fi-enabled devices are predicted to cross the 200 million mark within 4 years. Multiply times even a ten megabyte file, and they'd be passing almost 2 petabytes of data per day. I don't think they've thought this through. That's about a tenth the traffic of Google or Steam. They're either completely s**t-for-brains stupid or they're trying to guarantee that this fails....