I'll give you another one: Pro-life vs. pro-choice.
And that's the problem with creating Newspeak. No matter how many times you call yourself pro-life, your opponents are never going to call themselves anti-life. You'll almost never see a Democratic politician get on the television and denounce tax cuts, but you will see them advocating "revenue enhancements" or "making the rich pay their share" (meaning pay more, even though they already pay 90% of the taxes) etc.
Each side uses their own Orwellian phrasing. Nobody internalizes the other side's. The words meaning the opposite of the Newspeak words are still in the collective language and the concepts they represent can all still be articulated.
In other words, the concept you are referring to is not a subversion of language. It is far simpler: A free market and an unregulated market are not the same thing, but saying that they are is a lie, not a subversion of language.
I imagine they cost more to build in the first place. I don't see why it would cost anything more to operate: It's a giant radiator. It just sits there.
a giant atom bomb that's kept from exploding with a barrier
That's not how it works. There is never any exploding. A meltdown occurs if the reactor produces more heat than the cooling system can remove (for example, because the cooling system failed), and the reactor chamber temperature increases until things that were formerly solid start to melt, which ruins the reactor. Notwithstanding all the media hype, the consequences of a nuclear meltdown are almost exactly the same as the consequences of a building that stores spent fuel rods burning down after an electrical fire, except that the latter is actually worse (at least for US reactors) because the reactor proper would be inside of a containment building.
That means one organism per 2.5 * 365 = about 912 gallons. That can't be distinct algae cells unless its nearly sterile water.
You're doing the math assuming that every living organism is killed, and so the numbers can't be right for algae because there must be more algae than that. I suspect what is actually going on is that some very small minority of algae (and other organisms) are killed, but there is so much algae that it ends up being a big number. After all, who other than pedant scientists would call a fish an "organism"?
Are you sure they're (inherently) evaporative? There is no obvious reason that you can't create a closed coolant system that works like a larger version of what is used in a car or home radiator.
The ironic thing about this situation is that the entire problem could be solved (especially for newer reactors) by building cooling towers rather than using rivers for cooling. But cooling towers look scary, so nobody likes them.
I think Orwell had that much right. If you can control the vocabulary, you can control the discussion. If you control the discussion, you can control the conclusion.
What I'm saying is that you can't do that. The Party wants you to be able to say "The Party is good" and "Death to The Enemy," but if that vocabulary exists then you can use it to construct the sentence "Death to The Party" because language is combinatorial. And changing the name from Department of War to Department of Defense doesn't make people any less angry when you send their children to some miserable desert country to die for no apparent reason.
Changing the names of things is not totally without effect (that's why they do it), but the effect is always temporary. They call it the Department of Defense, people start calling it the DoD. And no matter what they call it, it quickly becomes an idiom rather than a phrase: Once people internalize that "the Department of Defense" means "the military," the specific words in the phrase are no longer parsed individually. Words become terms of art in different contexts. So you can create sentences like Yogi Berra, which you aren't "supposed" to be able to: "A nickel isn't worth a dime today." "Nobody goes there anymore because it’s too crowded." "The future ain't what it used to be."
Even in Orwell's own silly language, you can form the sentence "crimethink double plus good." And even if you aren't allowed to say it, you will always be able to think it.
The problem is that right after you don't buy into the hype (and expensive products), some less-than-cluefull employee will give out his/her password over the phone, or download and run some malicious attachment.
That is not really the problem. The problem is that too many congress critters subscribe to the Legislator's Fallacy: "Something must be done. This is something. Therefore, we must do this." If not for that, the existence of dim witted federal staffers could be resolved by firing them (or not hiring them in the first place) rather than spending a trillion dollars a year fighting an imagined enemy.
One of the things people have the hardest time accepting is that sometimes Bad Things Happen and the cost of preventing them exceeds the cost of allowing them to happen. In other cases the problem is a legitimate problem but the solution offered is totally irrational because the better solution requires goring the wrong constituency's ox, and with the rational solution taken off the table for political reasons, people are unhappy that the problem is not being solved and demand the outrageous and ineffective solution.
Of course, in this case it isn't really any of those things: This is just garden variety corruption. If you want to divert a trillion tax dollars into your own pocket then you need to pretend you're providing something of value to the general public. Saving them from imaginary cyber attacks (or whatever) is as good an excuse as any -- and hey, if there are no cyber attacks, it must mean they're doing their job. And if there are cyber attacks, it must mean they need more tax money.
Orwell tried to warn us. See also his work on the use of language and using it as an agent of control (Chomsky says basically the same thing).
Orwell got a lot of things right, but his arguments about use of language were pretty wrong. You can't actually create Newspeak. If you start calling copyright infringement piracy, people start to think that pirates are cool and piracy means sticking it to The Man. If you decide that calling it piracy is no longer cutting it and start calling it theft, people will soon start making references to Robin Hood instead of Captain Jack Sparrow. (You must admit that the pigopolists bear a closer resemblance to the Sheriff of Nottingham than they do to the British Navy.)
Chomsky has it more right, but despite being a linguist his points aren't as much about language as information: The issue is that selection bias allows you to tell part of the truth, and then defy anyone to prove that your biased selection is empirically false rather than merely intentionally incomplete, leaving the general public with the impression that the things the media says are irrefutable because no one is allowed any opportunity to refute them. In other words, the problem is not that powerful people choose what you are allowed to say or even how you are allowed to say it, it is that the content of your message determines how large of an audience you are allowed to reach.
That sounds suspiciously like elevating form over substance. You could infringe the second claim with a four function pocket calculator. Moreover, if you can take something unpatentable and add "but on a computer," when "a computer" is in the prior art and, given the unpatentable thing, doing it on a computer is totally obvious, that seems like more sleight of hand than reasoned argument: If the entirety of the novelty of the alleged invention is the unpatentable part of it, you can't very well issue the patent and consistently claim that the thing is unpatentable. If you aren't patenting the thing that supposedly isn't patentable then what novel and non-obvious thing are you patenting?
Is your point that Bilski made all software patents invalid? Because you can do anything in your head that you can do with pencil and paper if you take enough time to memorize it all, and you can do anything with pencil and paper that you can do with a computer.
Unless your point is that some programs are too complicated for someone to memorize during a normal human lifetime, and only those should be patentable? Because that seems like an extraordinarily vague and arbitrary line to draw.
Suppose that a company has several corporate executives who each pay 35% in taxes of every additional dollar they earn. The company also employs several line workers who pay no income tax because of the earned income tax credit. If the executives conclude that they can cut wages or staff at the factory and pay the savings to themselves as bonuses, Congress has a financial incentive to promote that behavior, because they get 35% of the executives' bonus money and don't lose any tax revenue from the line workers.
Congress doesn't even know they're doing this half the time. Every time a bill is proposed, the Congressional Budget Office writes a report about its impact on the federal revenues. If the bill will cause the rich to make less money (regardless of how much it helps the poor), it creates a bigger hole in the budget and the bill gets canned. If it will cause the rich to make more money (even if it is directly at the expense of the poor), it makes the hole in the budget smaller and anyone who tries to defeat the bill is charged with making up the revenue it was calculated to generate.
A basic income doesn't do that, because everyone gets the same basic income no matter how much money they make. There is no longer an incentive for the government to cause a thousand blue collar workers to lose $10,000/year in compensation so that the same money can be paid to Steve Ballmer or Brian Moynihan, because they would all pay the same marginal tax rates, even though the effective tax rate (after subtracting the basic income) for everyone would be approximately what it is now.
The only difference is ideology. The progressives want to make decisions for your own good, the conservatives just want to sell you out to big business.
I don't think that's strictly true. The progressives want to make decisions and tell you it's for your own good.
Give you an example: Progressives like the graduated income tax. If you actually do the math, it's a kind of a stupid thing: Poor people come out better off if you combine a uniform flat rate tax with a basic income. The uniform flat tax removes the incentive for Congress to adopt policies that divert income from the poor to the rich, who pay higher tax rates, so that Congress itself has more money to spend. At the same time, the basic income offsets any taxes paid by the poor and then some, and it removes the need for separate bureaucracies like unemployment insurance and social security.
But if you make the tax uniform then you remove the ability of the government to create subsidies for big business and call them tax cuts. You remove the ability of lobbyists to argue that regulatory capture bills that suck money out of the middle class and small business owners and into the pockets of big banks and the super rich will help close the deficit, because the latter no longer pay higher nominal tax rates and that sort of reverse redistribution of wealth no longer increases government revenues. If you create a basic income then you increase labor prices without the need for a union, because people no longer need to accept the first job offer they get at any wage in order to keep food on the table. The "progressives" don't actually want any of that -- those things are staples of the Congressional meal ticket. The "conservatives" don't want it either, for exactly the same reasons. Which is why it doesn't happen.
Moral equivalence doesn't really refute the moral panic argument... there is nothing less unsavory that the behavior in question is alleged to be equivalent to.
If anything it supports that this is a moral panic: Topless seventeen year olds are in no way equivalent to fornicating six year olds. Yet what we see consistently is that legitimate outrage directed at the latter is used to create draconian laws and massive public pressure to eliminate the former, even though it isn't in any way more harmful than if the same picture was taken a month later when the girl was eighteen.
It's the same thing we see with the copyright industries: Actual counterfeit drugs that kill patients are terrible but not common whereas noncommercial copyright infringement is common but not terrible. Actual child pornography is terrible but not common whereas "barely not legal" jailbait is common but not terrible. They aren't the same thing and deserve to be addressed separately rather than indiscriminately lumped together.
This is a straw man argument. There is also no chemical process that will allow me to check my email. That doesn't mean that chemical processes aren't processes. All this argument proves is that chemical processes are different than software processes because they deal with chemicals, rather than electrical signals. No kidding.
It has nothing to do with electrical signals. Anything you can do in software you can do (obviously more slowly) with pencil and paper. You can do it in your head.
Your comment about checking your email is just proving the point. You can construct a chemical process that will allow you to check your mail. Information can be encoded in chemical compounds (the one most people think of is DNA, but there are other possibilities). The encoding method can be devised so that reactions between compounds represent mathematical operations and the products of the reaction represent the outcome of the calculations. You can build a chemical computer and run any program that you could run on any other Turing machine, including an email client. We have not found that to be more efficient than microprocessors, so we don't do it. If you found a way to make it practical, you could patent the machine -- but you ought not to be able to patent the software that runs on it, because software is just a series of mathematical operations.
Actually, I would say that the process patents for chemical processes implies this. It certainly fits the criteria many here have posted of working with physical objects.
The phrase "necessary but not sufficient" comes to mind. You can encode information in physical things. That doesn't make the information patentable, and it only ought to make the physical thing patentable to the extent it would be patentable independent of what information is encoded.
No, I think the problem is rather that we haven't found a good way to make patents apply to programs running on general purpose computers.
I can tell you why that is: The two most striking problems with software patents are that they are granted far, far more broadly than they ever ought to be, and that independent invention is not a defense even though it occurs vastly more frequently in software than in other fields. Addressing those problems will make software patents look almost exactly like software copyright, at which point spending many thousands of dollars prosecuting a software patent will be wholly superfluous and they might as well not exist.
Why is patenting encryption by modular exponentiation absurd?
There is no such thing as "encryption by modular exponentiation" as distinct from "modular exponentiation." They are the same piece of code. You don't see why patenting "(x^y) (mod z)" is absurd? Start with "it's in the prior art."
I had no idea modular exponentiation could be applied to encryption. If nobody else either had noticed this - then why should we not allow somebody to protect their invention?
Well, we can start with "somebody else had noticed this." Clifford Cocks at GCHQ (UK intelligence) was the first to discover the relationship in 1973.
But never mind that. You're asking the wrong question. The question is, why should we set back the field of public key cryptography by 20 years because too many third parties are unwilling to adopt a patented algorithm in any kind of standard?
What are the alternatives?
1) The idea is never put into use. Humanity doesn't get encryption by modular exponentiation.
You have provided no suggestion as to why this would have happened. The developers of RSA worked for MIT. Their jobs were to publish papers to maintain the prestige of the school as a top research institute. MIT has close to a ten billion dollar endowment. The probability that they would do less research if it couldn't be patented is approximately zero.
2) The inventor builds an encryption product out of the invention. BigCorp Inc. reverses the code of the encryption program, and next version of their OS has encryption by modular exponentiation. The inventor gets a email saying "thanks for all the fish". Or a job offer, if lucky. 3) Inventor publishes the invention in a engineering journal. Same outcome as in case 2.
What you're describing could be better put as "the discovery is published and the world is better off because now everyone can use it." The reference to major corporations adopting it is a red herring to stir anti-corporate sentiment; the large majority of software patents are filed by major corporations.
Luckily the concept itself cannot be patented. But if I come up with a new use for modular exponentiation - don't you want me to share? Or do you just want me to share it for free. As in beer.
You'll share it because it's the only way for you to make use of it yourself, and you'll benefit because you'll have a first to market advantage, and because you can copyright the implementation which prevents anyone from copying the idea itself without developing and testing their own implementation. Since the implementation almost invariably costs more than the idea to develop, anyone who copies the idea is on the same footing with respect to fixed costs as the first mover and so is incapable of undercutting the first mover's prices.
Thus we can rephrase the patent argument as: should people be rewarded for being the first to think of something?
I think more importantly: Should people be punished for being the first to successfully implement something they thought of merely because they were not the first to think of it?
I'm not saying I like software patents, but I fail to see how a process that involves complicated chemistry is any less "math" than software.
Chemistry is not math. Chemistry is described by math. There is no set of equations or software program that will actually cause crude oil to become gasoline.
It seems that they're both simply algorithms. One deals with how you control chemicals, the other with how you control electrons.
By this logic you can patent a work of fiction as a process of arranging letters on a printing press plate.
The problem is that software is purely abstract. Nothing stops you from having a power source that can supply five times as much energy as the sun, or a gear with a hundred million teeth, or a totally frictionless surface. There is nothing useful you can add as a claim limitation to reduce the scope of the patent from an idea to an implementation, without giving potential infringers a trivial way to avoid the patent.
Example: Suppose you want to patent RSA encryption. Limiting yourself to specific numbers of bits in the key or using a modulus from a finite list enumerated in the patent or anything of that nature would only be an arbitrary limitation that would allow anyone to easily avoid infringing the patent. But the alternative is the absurdity of issuing a patent on modular exponentiation.
Consumers -- the people using the software -- are comparing desktops to desktops. It doesn't matter why your brakes failed.
That's my point: The fact that Windows Update works pretty well doesn't much matter if your computer gets infected with "XP Antivirus 2012" every other week.
I've literally never, not once, in my experience with my own machines, my co-workers' machines, or my friends' machines seen a Windows box be brought down by a video driver upgrade -- ATI, nVidia, whatever. But everyone I know who uses Linux (myself included) has found themselves doing "lynx google.com" after a video driver upgrade.
At this point we're just trading anecdotes, but I can tell you the exact scenario that causes the problem in Windows: Somebody has a newish machine and is trying to play a newish game, the game is running like crap or not running at all, and the solution proffered is to update the video driver. The video card being somewhat recent, the latest driver on the manufacturer's website has been available for a total of about 36 hours, and having installed it the next thing you see is a BSOD.
This is a more complex issue. See, I don't have to upgrade Windows to get newer versions of software. Modulo backporting trickery, that's not true of Ubuntu -- if I want a newer GCC or whatever I either a) build from source, destabilizing my packages or b) do a release upgrade, breaking my machine.
I don't know about that. Ever try installing IE9 on Windows XP? Or using any of that baleful Hollywood nonsense that requires the new-to-Vista DRM stack? Or DirectX 11?
Also, most software that you would want to upgrade outside of the standard release window will have a PPA available. For example, for gcc see here.
We suspect it was because of a regression with the new dm's compatibility with LDAP'd network shared home directories.
This is what I mean by a shared characteristic. Failures that all have the same cause aren't statistically independent. Your 70% number is meaningless if the issue only affects people who use LDAP with network shared home directories and the large majority of users don't do that.
Funny, the guys running Windows upstairs don't need a service history (to a rough approximation). Are you listening to yourself? You're seriously suggesting it's okay to have to document getting a consumer machine running. We document install procedures on supercluster nodes with exotic hardware. That shouldn't be necessary for desktops with commodity hardware.
And it isn't, by and large. You can pretty much take anything Dell has sold in the last decade and install Ubuntu 10.04 on it with no problems. You're the one claiming that some extravagant list of problems will be had in doing a clean install. That hasn't been my experience. What I'm saying is that if you're doing something exotic, or you run into some strange corner case that requires a non-obvious solution, it would be wise to document it -- if only so that when the machine gets hit by lightning you can get it back in working order. But once you've documented it, you don't have to worry about it again the next time you do a clean install, because you've now got the solution available should you need it.
The last 10% takes 50% or more of the work, and Canonical doesn't have anything like the funds to pay for that -- it's borderline dishonest to suggest otherwise.
I would agree with you if Canonical were the only ones doing the work, but they aren't. Linux is huge in the server space -- which means all of the problems with network drivers, file systems, multitasking, etc. get solved by those people, who are spending the same billions of dollars that Microsoft does. All Canonical has to do is take the fine work of those people and polish the UI.
Ships haven't gotten all that much faster over the years; modern container ships are only about twice as fast as the last clipper ships, not ten or a hundred or a thousand times as fast.
That's only because nobody cares about making them faster. They already run them slower than their maximum speed intentionally because the primary design consideration for container ships is efficiency, not speed. And if you look at the amount of freight transported per ship and per crew member compared to a clipper ship, I suspect it is in the range of ten to a hundred times more.
But let's assume you're right:
rockets are a mature technology, like ships and aircraft.
OK, so don't use rockets. I keep hearing about this magic carbon nanotube space elevator that we'll have Real Soon Now.
Even if you want to assume that never happens, let's consider another alternative: You pick a proverbial "asteroid the size of Texas" out of the many floating around out there. Find one in the habitable zone. Then you send a team there with some industrial equipment, not to mine the asteroid and bring it home, but to mine it and use the raw materials to construct a large compartmentalized living environment. Start off by building a few dozen acres of greenhouse, and fill it with plants so that you have food and oxygen recycling. Then proceed to build yourself a little home away from home -- but in a place that has lower gravity, so that you can build yourself a space elevator. Create yourself a city in space with a population of a few thousand people.
That gives you a foothold. You create an industrial city that can export its products to the universe without burning ten thousand gallons of fuel fighting gravity. Then you can branch out. Colonize and mine more asteroids and small planets. Once you have a large enough industrial capacity in low gravity areas, you can build yourself a city-to-go and just land the entire thing piecemeal on a suitable planet. Next thing you know you've got a million people living on Mars and several expeditions on their way to colonizing habitable planets in other planetary systems.
I'm not saying what I've just laid out would definitely work. Maybe, maybe not. What I'm saying is that we haven't yet exhausted all the possible alternatives, so giving up now is nothing but defeatism.
A billion other people don't actually have the money to fly a private jet anywhere.
And if you try to do something that would cause those billion other people to reduce their oil consumption, like raising the gas tax, they start a lynch mob and accuse you of not caring about poor people and destroying the economy.
The recent Ubuntu upgrades hosed about ten systems to the point where they wouldn't boot. I remember a Windows XP (IIRC) update years ago that had a chance of making a tiny percent of computers not boot, and people here went absolutely wild.
You're comparing apples and oranges. Windows update only applies critical updates to Windows. The Ubuntu package manager updates third party software, drivers, everything. I can't even count the number of times I've seen updating the video driver hose a Windows machine. Not to mention that the fact that common software comes as executable files from arbitrary websites encourages users to run stuff like that and hose their machine with "xp antivirus 2012" and the like.
And you're comparing upgrading from one release of Ubuntu to another with applying security patches through Windows Update. Have you ever tried upgrading from one version of Windows to another? It's such a well-renowned disaster that hardly anyone even attempts it. And if you were so bold as to skip installing Vista on your 2006-era machine that came with Windows XP so that you don't have to upgrade Windows once every three years, you'll find that an upgrade from Windows XP to Windows 7 is unsupported: You have to do a clean install and "take on every problem you ever had getting the box to run correctly."
But we're talking about something like a 70% failure rate with the Ubuntu upgrades.
I expect I'm right in assuming that the failures you experienced were related because the machines that failed shared a single characteristic that caused the failure. You would hardly have had a 70% failure rate if you had bothered to do a modicum of testing and determined the cause of the failure on the first machine that failed before wantonly breaking the rest of them.
In addition, "backing up/home and reinstalling" is not the simple process you make it out to be; you're taking on every problem you ever had getting the box to run correctly.
Those would be problems you've already solved. In the rare case that they haven't been fixed in the new release, you just pull the documented solution out of the service history and apply it.
On top of all of that, as of 12.04 due this April the Ubuntu LTS desktop releases will be supported for five years. Generally speaking the service lifetime of PC hardware is five years or less, so the idea that you'll have to worry about the consequences of upgrading from one release to another is about to become irrelevant because you'll have bought a new machine by the time the latest LTS release needs to be upgraded.
Ubuntu will never have the funding to pay people to fix the "last mile" of problems, because those aren't sexy.
This is what I'm talking about when I say cliched and outdated. Canonical is not a charity that fixes problems based on how sexy they are. They're a business that pays programmers to fix the unsexy problems their customers actually have.
Slashdot Mirror
I'll give you another one: Pro-life vs. pro-choice.
And that's the problem with creating Newspeak. No matter how many times you call yourself pro-life, your opponents are never going to call themselves anti-life. You'll almost never see a Democratic politician get on the television and denounce tax cuts, but you will see them advocating "revenue enhancements" or "making the rich pay their share" (meaning pay more, even though they already pay 90% of the taxes) etc.
Each side uses their own Orwellian phrasing. Nobody internalizes the other side's. The words meaning the opposite of the Newspeak words are still in the collective language and the concepts they represent can all still be articulated.
In other words, the concept you are referring to is not a subversion of language. It is far simpler: A free market and an unregulated market are not the same thing, but saying that they are is a lie, not a subversion of language.
I imagine they cost more to build in the first place. I don't see why it would cost anything more to operate: It's a giant radiator. It just sits there.
Do you have some evidence to support your theory that a nuclear power plant in Maine destroyed your local fishing hole?
That would be a pretty neat trick considering that there are no nuclear power plants in Maine.
a giant atom bomb that's kept from exploding with a barrier
That's not how it works. There is never any exploding. A meltdown occurs if the reactor produces more heat than the cooling system can remove (for example, because the cooling system failed), and the reactor chamber temperature increases until things that were formerly solid start to melt, which ruins the reactor. Notwithstanding all the media hype, the consequences of a nuclear meltdown are almost exactly the same as the consequences of a building that stores spent fuel rods burning down after an electrical fire, except that the latter is actually worse (at least for US reactors) because the reactor proper would be inside of a containment building.
That means one organism per 2.5 * 365 = about 912 gallons. That can't be distinct algae cells unless its nearly sterile water.
You're doing the math assuming that every living organism is killed, and so the numbers can't be right for algae because there must be more algae than that. I suspect what is actually going on is that some very small minority of algae (and other organisms) are killed, but there is so much algae that it ends up being a big number. After all, who other than pedant scientists would call a fish an "organism"?
Are you sure they're (inherently) evaporative? There is no obvious reason that you can't create a closed coolant system that works like a larger version of what is used in a car or home radiator.
The ironic thing about this situation is that the entire problem could be solved (especially for newer reactors) by building cooling towers rather than using rivers for cooling. But cooling towers look scary, so nobody likes them.
I think Orwell had that much right. If you can control the vocabulary, you can control the discussion. If you control the discussion, you can control the conclusion.
What I'm saying is that you can't do that. The Party wants you to be able to say "The Party is good" and "Death to The Enemy," but if that vocabulary exists then you can use it to construct the sentence "Death to The Party" because language is combinatorial. And changing the name from Department of War to Department of Defense doesn't make people any less angry when you send their children to some miserable desert country to die for no apparent reason.
Changing the names of things is not totally without effect (that's why they do it), but the effect is always temporary. They call it the Department of Defense, people start calling it the DoD. And no matter what they call it, it quickly becomes an idiom rather than a phrase: Once people internalize that "the Department of Defense" means "the military," the specific words in the phrase are no longer parsed individually. Words become terms of art in different contexts. So you can create sentences like Yogi Berra, which you aren't "supposed" to be able to: "A nickel isn't worth a dime today." "Nobody goes there anymore because it’s too crowded." "The future ain't what it used to be."
Even in Orwell's own silly language, you can form the sentence "crimethink double plus good." And even if you aren't allowed to say it, you will always be able to think it.
The problem is that right after you don't buy into the hype (and expensive products), some less-than-cluefull employee will give out his/her password over the phone, or download and run some malicious attachment.
That is not really the problem. The problem is that too many congress critters subscribe to the Legislator's Fallacy: "Something must be done. This is something. Therefore, we must do this." If not for that, the existence of dim witted federal staffers could be resolved by firing them (or not hiring them in the first place) rather than spending a trillion dollars a year fighting an imagined enemy.
One of the things people have the hardest time accepting is that sometimes Bad Things Happen and the cost of preventing them exceeds the cost of allowing them to happen. In other cases the problem is a legitimate problem but the solution offered is totally irrational because the better solution requires goring the wrong constituency's ox, and with the rational solution taken off the table for political reasons, people are unhappy that the problem is not being solved and demand the outrageous and ineffective solution.
Of course, in this case it isn't really any of those things: This is just garden variety corruption. If you want to divert a trillion tax dollars into your own pocket then you need to pretend you're providing something of value to the general public. Saving them from imaginary cyber attacks (or whatever) is as good an excuse as any -- and hey, if there are no cyber attacks, it must mean they're doing their job. And if there are cyber attacks, it must mean they need more tax money.
You clearly mean CFCTCFCY.
Orwell tried to warn us. See also his work on the use of language and using it as an agent of control (Chomsky says basically the same thing).
Orwell got a lot of things right, but his arguments about use of language were pretty wrong. You can't actually create Newspeak. If you start calling copyright infringement piracy, people start to think that pirates are cool and piracy means sticking it to The Man. If you decide that calling it piracy is no longer cutting it and start calling it theft, people will soon start making references to Robin Hood instead of Captain Jack Sparrow. (You must admit that the pigopolists bear a closer resemblance to the Sheriff of Nottingham than they do to the British Navy.)
Chomsky has it more right, but despite being a linguist his points aren't as much about language as information: The issue is that selection bias allows you to tell part of the truth, and then defy anyone to prove that your biased selection is empirically false rather than merely intentionally incomplete, leaving the general public with the impression that the things the media says are irrefutable because no one is allowed any opportunity to refute them. In other words, the problem is not that powerful people choose what you are allowed to say or even how you are allowed to say it, it is that the content of your message determines how large of an audience you are allowed to reach.
That sounds suspiciously like elevating form over substance. You could infringe the second claim with a four function pocket calculator. Moreover, if you can take something unpatentable and add "but on a computer," when "a computer" is in the prior art and, given the unpatentable thing, doing it on a computer is totally obvious, that seems like more sleight of hand than reasoned argument: If the entirety of the novelty of the alleged invention is the unpatentable part of it, you can't very well issue the patent and consistently claim that the thing is unpatentable. If you aren't patenting the thing that supposedly isn't patentable then what novel and non-obvious thing are you patenting?
Is your point that Bilski made all software patents invalid? Because you can do anything in your head that you can do with pencil and paper if you take enough time to memorize it all, and you can do anything with pencil and paper that you can do with a computer.
Unless your point is that some programs are too complicated for someone to memorize during a normal human lifetime, and only those should be patentable? Because that seems like an extraordinarily vague and arbitrary line to draw.
Suppose that a company has several corporate executives who each pay 35% in taxes of every additional dollar they earn. The company also employs several line workers who pay no income tax because of the earned income tax credit. If the executives conclude that they can cut wages or staff at the factory and pay the savings to themselves as bonuses, Congress has a financial incentive to promote that behavior, because they get 35% of the executives' bonus money and don't lose any tax revenue from the line workers.
Congress doesn't even know they're doing this half the time. Every time a bill is proposed, the Congressional Budget Office writes a report about its impact on the federal revenues. If the bill will cause the rich to make less money (regardless of how much it helps the poor), it creates a bigger hole in the budget and the bill gets canned. If it will cause the rich to make more money (even if it is directly at the expense of the poor), it makes the hole in the budget smaller and anyone who tries to defeat the bill is charged with making up the revenue it was calculated to generate.
A basic income doesn't do that, because everyone gets the same basic income no matter how much money they make. There is no longer an incentive for the government to cause a thousand blue collar workers to lose $10,000/year in compensation so that the same money can be paid to Steve Ballmer or Brian Moynihan, because they would all pay the same marginal tax rates, even though the effective tax rate (after subtracting the basic income) for everyone would be approximately what it is now.
The only difference is ideology. The progressives want to make decisions for your own good, the conservatives just want to sell you out to big business.
I don't think that's strictly true. The progressives want to make decisions and tell you it's for your own good.
Give you an example: Progressives like the graduated income tax. If you actually do the math, it's a kind of a stupid thing: Poor people come out better off if you combine a uniform flat rate tax with a basic income. The uniform flat tax removes the incentive for Congress to adopt policies that divert income from the poor to the rich, who pay higher tax rates, so that Congress itself has more money to spend. At the same time, the basic income offsets any taxes paid by the poor and then some, and it removes the need for separate bureaucracies like unemployment insurance and social security.
But if you make the tax uniform then you remove the ability of the government to create subsidies for big business and call them tax cuts. You remove the ability of lobbyists to argue that regulatory capture bills that suck money out of the middle class and small business owners and into the pockets of big banks and the super rich will help close the deficit, because the latter no longer pay higher nominal tax rates and that sort of reverse redistribution of wealth no longer increases government revenues. If you create a basic income then you increase labor prices without the need for a union, because people no longer need to accept the first job offer they get at any wage in order to keep food on the table. The "progressives" don't actually want any of that -- those things are staples of the Congressional meal ticket. The "conservatives" don't want it either, for exactly the same reasons. Which is why it doesn't happen.
Moral equivalence doesn't really refute the moral panic argument... there is nothing less unsavory that the behavior in question is alleged to be equivalent to.
If anything it supports that this is a moral panic: Topless seventeen year olds are in no way equivalent to fornicating six year olds. Yet what we see consistently is that legitimate outrage directed at the latter is used to create draconian laws and massive public pressure to eliminate the former, even though it isn't in any way more harmful than if the same picture was taken a month later when the girl was eighteen.
It's the same thing we see with the copyright industries: Actual counterfeit drugs that kill patients are terrible but not common whereas noncommercial copyright infringement is common but not terrible. Actual child pornography is terrible but not common whereas "barely not legal" jailbait is common but not terrible. They aren't the same thing and deserve to be addressed separately rather than indiscriminately lumped together.
This is a straw man argument. There is also no chemical process that will allow me to check my email. That doesn't mean that chemical processes aren't processes. All this argument proves is that chemical processes are different than software processes because they deal with chemicals, rather than electrical signals. No kidding.
It has nothing to do with electrical signals. Anything you can do in software you can do (obviously more slowly) with pencil and paper. You can do it in your head.
Your comment about checking your email is just proving the point. You can construct a chemical process that will allow you to check your mail. Information can be encoded in chemical compounds (the one most people think of is DNA, but there are other possibilities). The encoding method can be devised so that reactions between compounds represent mathematical operations and the products of the reaction represent the outcome of the calculations. You can build a chemical computer and run any program that you could run on any other Turing machine, including an email client. We have not found that to be more efficient than microprocessors, so we don't do it. If you found a way to make it practical, you could patent the machine -- but you ought not to be able to patent the software that runs on it, because software is just a series of mathematical operations.
Actually, I would say that the process patents for chemical processes implies this. It certainly fits the criteria many here have posted of working with physical objects.
The phrase "necessary but not sufficient" comes to mind. You can encode information in physical things. That doesn't make the information patentable, and it only ought to make the physical thing patentable to the extent it would be patentable independent of what information is encoded.
No, I think the problem is rather that we haven't found a good way to make patents apply to programs running on general purpose computers.
I can tell you why that is: The two most striking problems with software patents are that they are granted far, far more broadly than they ever ought to be, and that independent invention is not a defense even though it occurs vastly more frequently in software than in other fields. Addressing those problems will make software patents look almost exactly like software copyright, at which point spending many thousands of dollars prosecuting a software patent will be wholly superfluous and they might as well not exist.
Why is patenting encryption by modular exponentiation absurd?
There is no such thing as "encryption by modular exponentiation" as distinct from "modular exponentiation." They are the same piece of code. You don't see why patenting "(x^y) (mod z)" is absurd? Start with "it's in the prior art."
I had no idea modular exponentiation could be applied to encryption. If nobody else either had noticed this - then why should we not allow somebody to protect their invention?
Well, we can start with "somebody else had noticed this." Clifford Cocks at GCHQ (UK intelligence) was the first to discover the relationship in 1973.
But never mind that. You're asking the wrong question. The question is, why should we set back the field of public key cryptography by 20 years because too many third parties are unwilling to adopt a patented algorithm in any kind of standard?
What are the alternatives?
1) The idea is never put into use. Humanity doesn't get encryption by modular exponentiation.
You have provided no suggestion as to why this would have happened. The developers of RSA worked for MIT. Their jobs were to publish papers to maintain the prestige of the school as a top research institute. MIT has close to a ten billion dollar endowment. The probability that they would do less research if it couldn't be patented is approximately zero.
2) The inventor builds an encryption product out of the invention. BigCorp Inc. reverses the code of the encryption program, and next version of their OS has encryption by modular exponentiation. The inventor gets a email saying "thanks for all the fish". Or a job offer, if lucky.
3) Inventor publishes the invention in a engineering journal. Same outcome as in case 2.
What you're describing could be better put as "the discovery is published and the world is better off because now everyone can use it." The reference to major corporations adopting it is a red herring to stir anti-corporate sentiment; the large majority of software patents are filed by major corporations.
Luckily the concept itself cannot be patented. But if I come up with a new use for modular exponentiation - don't you want me to share? Or do you just want me to share it for free. As in beer.
You'll share it because it's the only way for you to make use of it yourself, and you'll benefit because you'll have a first to market advantage, and because you can copyright the implementation which prevents anyone from copying the idea itself without developing and testing their own implementation. Since the implementation almost invariably costs more than the idea to develop, anyone who copies the idea is on the same footing with respect to fixed costs as the first mover and so is incapable of undercutting the first mover's prices.
I'm well aware of that. I'm using it as an example of a patent that should not have been issued (because it's a software patent).
Thus we can rephrase the patent argument as: should people be rewarded for being the first to think of something?
I think more importantly: Should people be punished for being the first to successfully implement something they thought of merely because they were not the first to think of it?
I'm not saying I like software patents, but I fail to see how a process that involves complicated chemistry is any less "math" than software.
Chemistry is not math. Chemistry is described by math. There is no set of equations or software program that will actually cause crude oil to become gasoline.
It seems that they're both simply algorithms. One deals with how you control chemicals, the other with how you control electrons.
By this logic you can patent a work of fiction as a process of arranging letters on a printing press plate.
The problem is that software is purely abstract. Nothing stops you from having a power source that can supply five times as much energy as the sun, or a gear with a hundred million teeth, or a totally frictionless surface. There is nothing useful you can add as a claim limitation to reduce the scope of the patent from an idea to an implementation, without giving potential infringers a trivial way to avoid the patent.
Example: Suppose you want to patent RSA encryption. Limiting yourself to specific numbers of bits in the key or using a modulus from a finite list enumerated in the patent or anything of that nature would only be an arbitrary limitation that would allow anyone to easily avoid infringing the patent. But the alternative is the absurdity of issuing a patent on modular exponentiation.
Consumers -- the people using the software -- are comparing desktops to desktops. It doesn't matter why your brakes failed.
That's my point: The fact that Windows Update works pretty well doesn't much matter if your computer gets infected with "XP Antivirus 2012" every other week.
I've literally never, not once, in my experience with my own machines, my co-workers' machines, or my friends' machines seen a Windows box be brought down by a video driver upgrade -- ATI, nVidia, whatever. But everyone I know who uses Linux (myself included) has found themselves doing "lynx google.com" after a video driver upgrade.
At this point we're just trading anecdotes, but I can tell you the exact scenario that causes the problem in Windows: Somebody has a newish machine and is trying to play a newish game, the game is running like crap or not running at all, and the solution proffered is to update the video driver. The video card being somewhat recent, the latest driver on the manufacturer's website has been available for a total of about 36 hours, and having installed it the next thing you see is a BSOD.
This is a more complex issue. See, I don't have to upgrade Windows to get newer versions of software. Modulo backporting trickery, that's not true of Ubuntu -- if I want a newer GCC or whatever I either a) build from source, destabilizing my packages or b) do a release upgrade, breaking my machine.
I don't know about that. Ever try installing IE9 on Windows XP? Or using any of that baleful Hollywood nonsense that requires the new-to-Vista DRM stack? Or DirectX 11?
Also, most software that you would want to upgrade outside of the standard release window will have a PPA available. For example, for gcc see here.
We suspect it was because of a regression with the new dm's compatibility with LDAP'd network shared home directories.
This is what I mean by a shared characteristic. Failures that all have the same cause aren't statistically independent. Your 70% number is meaningless if the issue only affects people who use LDAP with network shared home directories and the large majority of users don't do that.
Funny, the guys running Windows upstairs don't need a service history (to a rough approximation). Are you listening to yourself? You're seriously suggesting it's okay to have to document getting a consumer machine running. We document install procedures on supercluster nodes with exotic hardware. That shouldn't be necessary for desktops with commodity hardware.
And it isn't, by and large. You can pretty much take anything Dell has sold in the last decade and install Ubuntu 10.04 on it with no problems. You're the one claiming that some extravagant list of problems will be had in doing a clean install. That hasn't been my experience. What I'm saying is that if you're doing something exotic, or you run into some strange corner case that requires a non-obvious solution, it would be wise to document it -- if only so that when the machine gets hit by lightning you can get it back in working order. But once you've documented it, you don't have to worry about it again the next time you do a clean install, because you've now got the solution available should you need it.
The last 10% takes 50% or more of the work, and Canonical doesn't have anything like the funds to pay for that -- it's borderline dishonest to suggest otherwise.
I would agree with you if Canonical were the only ones doing the work, but they aren't. Linux is huge in the server space -- which means all of the problems with network drivers, file systems, multitasking, etc. get solved by those people, who are spending the same billions of dollars that Microsoft does. All Canonical has to do is take the fine work of those people and polish the UI.
Ships haven't gotten all that much faster over the years; modern container ships are only about twice as fast as the last clipper ships, not ten or a hundred or a thousand times as fast.
That's only because nobody cares about making them faster. They already run them slower than their maximum speed intentionally because the primary design consideration for container ships is efficiency, not speed. And if you look at the amount of freight transported per ship and per crew member compared to a clipper ship, I suspect it is in the range of ten to a hundred times more.
But let's assume you're right:
rockets are a mature technology, like ships and aircraft.
OK, so don't use rockets. I keep hearing about this magic carbon nanotube space elevator that we'll have Real Soon Now.
Even if you want to assume that never happens, let's consider another alternative: You pick a proverbial "asteroid the size of Texas" out of the many floating around out there. Find one in the habitable zone. Then you send a team there with some industrial equipment, not to mine the asteroid and bring it home, but to mine it and use the raw materials to construct a large compartmentalized living environment. Start off by building a few dozen acres of greenhouse, and fill it with plants so that you have food and oxygen recycling. Then proceed to build yourself a little home away from home -- but in a place that has lower gravity, so that you can build yourself a space elevator. Create yourself a city in space with a population of a few thousand people.
That gives you a foothold. You create an industrial city that can export its products to the universe without burning ten thousand gallons of fuel fighting gravity. Then you can branch out. Colonize and mine more asteroids and small planets. Once you have a large enough industrial capacity in low gravity areas, you can build yourself a city-to-go and just land the entire thing piecemeal on a suitable planet. Next thing you know you've got a million people living on Mars and several expeditions on their way to colonizing habitable planets in other planetary systems.
I'm not saying what I've just laid out would definitely work. Maybe, maybe not. What I'm saying is that we haven't yet exhausted all the possible alternatives, so giving up now is nothing but defeatism.
That's what 1 billion other people say, too.
A billion other people don't actually have the money to fly a private jet anywhere.
And if you try to do something that would cause those billion other people to reduce their oil consumption, like raising the gas tax, they start a lynch mob and accuse you of not caring about poor people and destroying the economy.
The recent Ubuntu upgrades hosed about ten systems to the point where they wouldn't boot. I remember a Windows XP (IIRC) update years ago that had a chance of making a tiny percent of computers not boot, and people here went absolutely wild.
You're comparing apples and oranges. Windows update only applies critical updates to Windows. The Ubuntu package manager updates third party software, drivers, everything. I can't even count the number of times I've seen updating the video driver hose a Windows machine. Not to mention that the fact that common software comes as executable files from arbitrary websites encourages users to run stuff like that and hose their machine with "xp antivirus 2012" and the like.
And you're comparing upgrading from one release of Ubuntu to another with applying security patches through Windows Update. Have you ever tried upgrading from one version of Windows to another? It's such a well-renowned disaster that hardly anyone even attempts it. And if you were so bold as to skip installing Vista on your 2006-era machine that came with Windows XP so that you don't have to upgrade Windows once every three years, you'll find that an upgrade from Windows XP to Windows 7 is unsupported: You have to do a clean install and "take on every problem you ever had getting the box to run correctly."
But we're talking about something like a 70% failure rate with the Ubuntu upgrades.
I expect I'm right in assuming that the failures you experienced were related because the machines that failed shared a single characteristic that caused the failure. You would hardly have had a 70% failure rate if you had bothered to do a modicum of testing and determined the cause of the failure on the first machine that failed before wantonly breaking the rest of them.
In addition, "backing up /home and reinstalling" is not the simple process you make it out to be; you're taking on every problem you ever had getting the box to run correctly.
Those would be problems you've already solved. In the rare case that they haven't been fixed in the new release, you just pull the documented solution out of the service history and apply it.
On top of all of that, as of 12.04 due this April the Ubuntu LTS desktop releases will be supported for five years. Generally speaking the service lifetime of PC hardware is five years or less, so the idea that you'll have to worry about the consequences of upgrading from one release to another is about to become irrelevant because you'll have bought a new machine by the time the latest LTS release needs to be upgraded.
Ubuntu will never have the funding to pay people to fix the "last mile" of problems, because those aren't sexy.
This is what I'm talking about when I say cliched and outdated. Canonical is not a charity that fixes problems based on how sexy they are. They're a business that pays programmers to fix the unsexy problems their customers actually have.