What closed source hobbyists? If you're not distributing it then there are no customers for you to be liable to. If you're distributing it for money or commercial advantage then you're a commercial software developer, not a hobbyist. And if you're neither of those things then what reason do you have not to release the source with the binary?
If you read the proposal, it isn't likely to harm hobbyists much -- it basically exempts open source from the requirements. The likely result will be for proprietary software companies to either go out of business and become service companies, and for software to be developed in the future as open source by hardware and service companies.
And what does that have to do with anything? If Google can pay $X and get people to install a browser that doesn't have its default home page and search engine periodically reverted to Microsoft, and that makes them $Y/year and Y > X, they're going to do it. Apparently that is the case, so here we are.
It seems to me the trouble is that it teaches followers to have unquestioning faith, even in the face of reason and evidence to the contrary. And then the text which further instructs them what to have absolute faith in is not absolutely free from ambiguity and internal contradiction, which allows malicious persons to convincingly argue that it supports their ends and gather unwarranted support for unconscionable acts from the faithful.
It would be theoretically possible to construct a system of morality free from ambiguity, but nobody would like the result. It would either have to be a decision tree including every possible choice under every possible set of circumstances that anyone could possibly ever encounter, and then specifying the morally correct action under those circumstances, or it would be unambiguous by having a disagreeable level of over-breadth in certain areas, e.g. "all homicide is wrong, no exceptions" -- because the satisfactory exceptions would turn it into the comprehensive decision tree. And naturally the decision tree, while possible in theory, fails in practice because no one can even predict much less enumerate every possible set of circumstances in order to specify ex ante the correct choice in each. To say nothing of the impossibility of internalizing such a thing in a way that would allow it to be used for everyday decisions.
Given that fact, any system of morality that requires adherents to have unquestioning faith in a necessarily imperfect belief system is fundamentally flawed.
Science makes no claim to be a system of morality. It is a system for determining "what is" rather than for determining "what should." People then ascribe meaning to that information according to their independent system of morality.
Religion very much claims to be a system of morality. And for most major religions, a frightening and dangerous one.
I do believe the level of discussion between the legal departments of the two companies would have been a bit more specific than the simplistic and vague statements that get posted around here.
The problem is that it usually goes something like this: We have several hundred patents which have some vague relation to what you're doing. Some of them you don't really infringe, others would probably be invalidated if challenged, but if you wanted to actually challenge them all you'd end up spending many millions of dollars. In the meantime the lawsuit makes the news, your stock price goes down and customers and developers shy away from your products. How about you just pay us the money instead?
Licensing one patent is patent licensing. Licensing dozens or hundreds of patents is a protection racket -- because the cost of actually analyzing and challenging them exceeds the cost of the protection money.
Clearly hardware is patent-able. So a novel arrangement of logic gates soldered onto a PCB should be patent-able, therefore the same configuration of logic gates implemented on a FPGA should be covered by that patent (it's the same logic gate arrangement after all)
You've already failed. Let's try it this way: Clearly hardware is patentable. So a novel arrangement of letters in movable type affixed to a printing press should be patentable, therefore the same story on a computer should be covered by that patent (it's the same text after all).
But obviously it isn't. You can't patent literature, in the same way as you ought not to patent software.
In the Nokia case Apple felt their IP was being undervalued and Nokia's overvalued. Cross-licensing is not automatically fair. Saying you hav to let us copy all your products if you want to use the 3G standard, also not fair.
But saying that you have to let us copy all your technologies if you want to use rounded corners or a touch screen is fair?
If Apple thinks the exchange is unfair then they're free to make products with rounded corners and no 3G and require Samsung to make products with 3G and no rounded corners and see which one succeeds in the marketplace, but you don't get to be a hypocrite and say that competitors have to license their stuff to us but we won't reciprocate.
Mutually assured destruction, obviously. If you sue someone, they sue you back, you both spend millions on litigation and end up worse off than if you had just negotiated a cross-license or reasonable royalties in the first place.
By contrast, Apple asked for this by firing the first shot.
Scenario: Your operative is in an unsecured location preparing for a mission. There is no SCIF in his vicinity. You learn new information which is relevant and must be communicated to him immediately.
It seems obvious that having a communications device which is as secure as practicable under those conditions is preferable to e.g. sending a completely unencrypted text message to his COTS cell phone.
Did you miss the "supermegacorp buried in bullshit" part of my theory?
I did not. The trouble is that that isn't what will happen. Nobody is really going to get a permanent injunction against the iPhone and then try to enforce it permanently. Especially not bidirectionally -- if you have an injunction against them and they get an injunction against you then you're both better off to sit down at the negotiating table. It's all just leverage in negotiating royalties or cross-licensing agreements. And it's largely a zero-sum game -- for every loser there is a winner who will be there lobbying Congress to keep their advantage, whether it's a megacorp crushing its smaller competitors or a patent troll putting a private tax on digital products and putting the money towards its lobbying efforts.
After all with their money they can crush the competition in a myriad of other ways besides patents and copyrights. Collusion and bribery for example.
Collusion doesn't work in a free market with low barriers to entry, which is what software would be without software patents. Bribery doesn't work without collusion because your competitors just pay bribes the other way and you don't get results. And both collusion and bribery are illegal.
Yeah, the NSA has a different security model than Apple.
For one thing, if the thing is really secure, it shouldn't matter that nefarious people get access to one -- that is one of the main things you need it to be secure against.
Of course, the way you do this is pretty obvious. You put plenty of memory in it but only read-only permanent storage which holds the OS and the device's unique private key, and store all other data "in the cloud" (i.e. on the NSA's secure server). You put a hardware AES engine on the CPU and have it encrypt everything in RAM. You have it establish an encrypted tunnel at all times to a secure building in spy central somewhere and send all other communications through that. Then you use two or three factor authentication to unlock the phone, which authenticates against the central server, and when the phone is locked the encryption key to decrypt most of memory is stored in the central location rather than on the phone. If the phone gets lost you disable its account on the server and it's instantly bricked because it can't even read its own memory, and it doesn't contain any sensitive data in permanent storage.
The ONLY way we are gonna get any kind of reforms after citizen united is if supermegacorp gets so buried in patent bullshit that they can't even release a single product here for having it tied up in court for a decade.
There is a major flaw in this strategy, which is that parasites are sticky.
Let's say we keep issuing ever more bullshit patents. Some coven of asshats the likes of Intellectual Ventures will acquire a few thousand of them and then extract several hundred million dollars each in settlements and lawsuits from shaking down the companies that actually make stuff. So then they've got a fat wallet and a proven business model -- what makes you think they aren't going to spend a billion dollars a year lobbying Congress to make sure it stays that way?
On top of that, you think the foreign manufacturers are going to put pressure on fixing the system, but it's completely the opposite. If the foreign companies are ignoring the patent trolls in their home countries, they can't sell the same products to the US market because they'll get sued the same as anybody else. So what happens? Market division. The US companies can't compete in foreign markets anymore because they have to abide the ridiculous US patent system, and the foreign companies stay out of the US because they don't want to have to either. Now ask yourself what incentive that drives for US companies -- if they fix the US patent system then foreign competition that had been avoiding the US in order to avoid litigation-induced bankruptcy will enter the market and compete with them. By contrast, if they can export the damage of software patents through treaties or foreign lobbying then they can use the warchests they originally built to balance one another to destroy their foreign competition.
Parasites are sticky and thrive on inefficiency. If you have fast-paced competition to be first to market, the parasites can't survive. If you let the market converge into a collection of oligopolies who stamp out innovative new competitors through litigation and resign themselves to paying the occasional tithe to a patent troll out of their monopoly rents, there could be nobody left without a vested interest in the status quo.
Didn't Handspring--now defunct--already do this like 12 years ago? There were all sorts of devices for their PDAS, including a GSM module.
When has that stopped anybody before? The lawyers will just try to find one little thing which is different and claim that it's new and patentable.
A decade ago we got a million patents that said something like "that thing everybody knows about, but on the internet." Now we're going to get a million patents that say something like "that thing everybody knows about, but on a mobile device." Because hey, adding that extra element lets you try to claim that nobody has ever done it before. And the patent office is profoundly lax in issuing rejections for obviousness.
This is a case in point: Making a computing device modular? Who could have ever thought of that? (People in the 1950's, that's who.) But you throw in "it's a phone too" and you've got yourself a patent application, apparently.
Regardless of the holding duration, that's a buyer.
Not if they already have the stock and are deciding when to sell it.
This is not a valid premise
The premise is not required, it is only a simplification to make the explanation easier. Let me put it a different way: Suppose that in the world without market makers, the majority of people would expect it to go up by approximately the specified amount. I am only trying to create a baseline in order to analyze the effects of modifying it.
Your further analysis basically comes to the conclusion that market makers remove capital from the system, depressing stock prices. Among other things, this ignores the set of participants who lose money on their trades.
Unless you are trying to argue that market makers on average have non-positive net profits, I'm not sure what this is supposed to be buying you.
Furthermore, anyone who doesn't want to pay for immediacy doesn't have to. You can submit a limit order for a specific price, and wait for the normal volatility of the market to come to your limit price.
That is something the seller can do regardless of the effect I'm talking about, so I'm not sure how it is relevant.
You are imagining the "penny shaving" from a buyer's perspective, but market makers view buying and selling as mostly symmetric (buy low, sell high; or sell high, buy low), hence why the spread is impacted and not the "price". In actuality, HF market makers are competing with each other for increasingly small penny-shavings, and they're able to do so because reduced latency leads to reduced risk.
I am imagining it from the perspective of anyone who is not a "market maker" -- someone who intends to hold the stock for more than a minute.
Let's take an example. A company has 1000 shares of stock. 800 are held by long-term investors who all intend to keep them for at least a year as long as nothing drastic happens. 200 are held by short term investors, who each intend to sell them within a short period of time X. The buyers for those 200 shares in turn intend to hold them each for another X period and then sell them again to another short-term investor, and so on indefinitely.
Now let's say that the "true" value of the company will increase by 2 cents/share every period, and everybody knows and expects that. Moreover, the short-term investors demand that profit of 2 cents per share every period at the current stock price, otherwise they'll take their money elsewhere and buy bonds or commodities or something. If there are no "market makers" then every period the short-term investors sell their stock for 2 cents/share more than they bought it for previously. Now enter the market makers. Let's say that for each trade, the market makers take an average of 1 cent/share.
Look at the chain it creates: The prospective buyers in the first period demand a 2c/share profit over the next period; they won't buy the shares for any more than what is necessary to achieve that. They know the market makers will take 1c, so what they demand is a 3c increase in the bid price over the period that they hold the shares. Based on the fundamentals the price will only go up 2c however. No one will pay more than the share is worth at the end of the period, so to create the required difference the immediate share price must go down by 1c.
But wait, there's more: The same dynamic will exist in the next period: The buyer in period two demands a 2c profit, which requires a 3c bid price increase over one period, and the natural value will again only increase 2c. In consequence the bid price for the stock in period two will be reduced by 1c/share. The buyer in period one realizes this, and still demanding a 2c profit, will now bid 2c/share less than due to the market makers -- 1c to cover the market maker's cut on this trade and 1c to cover the amount the next bidder demands to cover the market maker's cut on the next trade. And it goes on like this, driving down the current-day price of the stock because short-term investors are willing to pay less and less for it because the HFT market makers are over time predicted to consume that portion of their expected profits.
Regardless, a long-term investor is looking profit at least an order of magnitude above the cost of liquidity, so none of this should matter to Joe Sixpack and his 401k.
So you're looking at it from the perspective of the long-term investor and saying that it doesn't matter to them because they're going to buy for $N/share and sell for $2N/share, and the penny or two per share shaved off at the back by a market maker doesn't matter one lick to him because it's 0.5% of the profit. The trouble is that most of the buyers on the market at any given time are the short-term investors. Long-term investors go years without buying a share; they are not regularly available as buyers to other long-term investors. So when it comes time for a long-term investor to sell his shares, the value of the shares to the short-term investors is what matters. If the market makers materially decrease the value of the shares to those short term investors, it will correspondingly impact the price at which the long-term
It's possible that HFT is contributing to it to some small extent, but I would think that other factors would have more to do with it. For one thing, the super rich have highly diverse investments. They own stocks, but also government bonds, corporate bonds, land, commodities, etc. It is very rare that all of those things go down together, and in fact usually when one of them goes down another goes up. They also have a bunch of high-priced investment advisers whose entire job it is to make sure they aren't on the losing side of the deal and that if they might be then it's properly hedged.
On top of that, a lot of these people outright own a large private company, and a lot of those companies will regularly beat the market because they're run by the owners rather than officers overly focused on quarterly profits and afflicted with the principle-agent problem.
The thing that concerns me about HFT is that it's quite possibly a significant cause for why the stock market is doing so poorly. Think about it: If the HFTers shave a penny or so per share off of each trade between exchanges, future prospective buyers will be willing to pay that much less for their shares, because they know that when they ultimately sell them they'll suffer the same loss again to the HFTers. In consequence the seller is obligated to lower the price by that small pittance per share in order to make the trade, which sets the new market price for the stock. The value of the stock will continue to go down until it accounts for the amount of value the HFTers remove over the period of time that the average investor holds the stock.
The more the HFTers make over a given period of time, the less the stock is worth to "real" investors. More to the point, if the "true" value of the company does not increase faster than the speed at which HFTers remove value from stockholders through arbitrage, the value (and therefore stock price) of the company will continue slowly falling indefinitely, because no one will want to buy shares that will post-arbitrage be worth less than they paid for them. (In practice what will happen is that the value will fall until the only remaining investors holding the stock will be long-term investors, who reduce the value removed by HFTers because they no longer have a sufficient volume of trading to arbitrage, and so the stock will stabilize at a non-zero price where the value of the company is increasing faster than the arbitrage removes, but at a much lower value than it would be absent HFT.)
That isn't really much protection if you consider what drivers are. By their nature they have to have privileged access to hardware. So you don't have kernel mode yet, but you're a driver? You put some rootkit code in a noncritical register or buffer on whatever piece of hardware the driver is operating, then you instruct the device to write that data to a physical memory address where the kernel will eventually execute it in kernel mode.
And that's ignoring the "most" -- if some drivers are not virtualized then all you have to do is have the click-happy user install the rootkit purveyor's own unsigned driver, which is one of the sort that has to run in kernel mode.
But that's not caching at all, it's prefetch. The filesystem driver or the program itself can do that completely independent of the underlying cache architecture simply by reading the data in anticipation of it being needed soon.
And prefetch is no panacea. It makes huge mistakes. If you're playing that same game and there are 10GB worth of 300 different levels and it tries to load them into the cache when you're loading the first one, you're creating a bunch of unnecessary disk activity (and wearing out the SSD), and pushing things out of the cache that have actually been used recently, even though the user is very unlikely to want to use all 300 levels during one session.
Is it aware of the filesystem and keeps most accessed files, or just the most accessed blocks?
How is that an advantage? If I have e.g. a 5GB mail archive file which is my most accessed file, but I really only frequently access a few dozen MB of it which represent the most recently received mail, the performance will be much better if it will cache the frequently accessed blocks only and then use the rest of the SSD to cache other recently accessed blocks than if it uses up half the SSD to cache the whole mail archive even though I haven't accessed 7/8ths of it in the last two years.
People are having trouble understanding relativity. That is not an alternative.
Scenario 1: Apple continues selling iPhone 4 and does not release iPhone 5. Apple captures only 10% of the market, Android captures 60%, non-Android competitors capture 30%. If Android did not exist, Apple would capture 35% of the market and non-Android competitors would capture 65%.
Scenario 2: Apple brings out the iPhone 5. Apple captures 20% of the market, Android captures 55%, non-Android competitors capture 20%. If Android did not exist, Apple would capture 55% of the market and non-Android competitors would capture 45%.
In any scenario, the existence of additional competition reduces market share and/or margins that the other competitors would otherwise have.
The problem for Apple is that maintaining high margins by continually improving your product in a competitive market becomes a losing battle against diminishing returns. R&D effort into a new product line will sensibly go after the low-hanging fruit first, but once that is gone and all competitors have those features as a new baseline, it becomes more and more difficult as time passes to collect higher margins than competitors by releasing incremental iterations of existing products. The old products achieve most of the features that most people want, and the number of people willing to pay a $200 or $300 premium for version N+1 diminishes.
In order to maintain consistent profitability with tech products (as distinguished, in particular, from services), a company without an intrenched monopoly is required to continually enter or create new markets, as low-cost competitors naturally enter existing markets and drive down profits. Apple has had some success with this by getting into music players, then phones, now tablets. The question is, can they keep it up long-term? And the answer to that question is by no means clear.
Not really. Just because a signed kernel was loaded initially doesn't mean that's what is still running by the time the anti-malware program gets running. As soon as you have something malicious running with the ability to execute all privileged hardware instructions, there is no way to rely on anything anymore. The rootkit can just load a whole new kernel if it has to, or it can patch the running kernel in memory. The best anti-malware can do is to detect it before it executes in kernel mode and prevent it from executing. And trying to get the anti-malware running as early as possible only makes you more susceptible to rogue anti-malware.
1) Corporations would naturally have a BIOS password which only the IT department would know, and they would know better than to type it just because the dancing bunnies "for some reason" require you to install a boot loader.
2) If you can convince the user to reboot the computer, enter the BIOS and change the authorized boot loader by hand in order to see some lolcats, there is just nothing you can do to overcome that level of stupidity. You might as well worry about a denial of service attack whereby the website instructs the user to stick a fork in the electrical outlet. If the user is that stupid, you lose, and you deserve to lose.
Gates felt that Linux devs gained some sort of undue enrichment. It's a bogus and douchbag position, but it's not "rent seeking".
You're reading his bogus defense of his conduct and ignoring the conduct. Copyright and patent holders do this on a regular basis. They say that they created something and demand remuneration, but what they demand is far in excess of what they contributed. The patent holder patents the parts to his copy machine, but then tries to leverage the "legitimate" patent monopoly over replacement parts into a monopoly over copier service. The motion picture industry takes their oligopoly position in copyrighted motion pictures and tries to leverage it into control over the distribution channels, and then over all consumer electronics.
If Microsoft breaks ACPI for Linux, they break more than they built. They are (to use the Microsoft camp's philosophy and terminology) "stealing" the benefit to the hardware makers who developed ACPI of selling ACPI-functional hardware to users of non-Microsoft operating systems. They are also ignoring the legitimate work done by the Linux camp to make ACPI work on Linux, as though somehow only Microsoft's efforts to make ACPI work for Windows are legitimate and require compensation and consideration but any efforts by a third party to make it work for another operating system can be ignored in the calculations.
Slashdot Mirror
What closed source hobbyists? If you're not distributing it then there are no customers for you to be liable to. If you're distributing it for money or commercial advantage then you're a commercial software developer, not a hobbyist. And if you're neither of those things then what reason do you have not to release the source with the binary?
If you read the proposal, it isn't likely to harm hobbyists much -- it basically exempts open source from the requirements. The likely result will be for proprietary software companies to either go out of business and become service companies, and for software to be developed in the future as open source by hardware and service companies.
Which I suppose wouldn't be the end of the world.
And what does that have to do with anything? If Google can pay $X and get people to install a browser that doesn't have its default home page and search engine periodically reverted to Microsoft, and that makes them $Y/year and Y > X, they're going to do it. Apparently that is the case, so here we are.
It seems to me the trouble is that it teaches followers to have unquestioning faith, even in the face of reason and evidence to the contrary. And then the text which further instructs them what to have absolute faith in is not absolutely free from ambiguity and internal contradiction, which allows malicious persons to convincingly argue that it supports their ends and gather unwarranted support for unconscionable acts from the faithful.
It would be theoretically possible to construct a system of morality free from ambiguity, but nobody would like the result. It would either have to be a decision tree including every possible choice under every possible set of circumstances that anyone could possibly ever encounter, and then specifying the morally correct action under those circumstances, or it would be unambiguous by having a disagreeable level of over-breadth in certain areas, e.g. "all homicide is wrong, no exceptions" -- because the satisfactory exceptions would turn it into the comprehensive decision tree. And naturally the decision tree, while possible in theory, fails in practice because no one can even predict much less enumerate every possible set of circumstances in order to specify ex ante the correct choice in each. To say nothing of the impossibility of internalizing such a thing in a way that would allow it to be used for everyday decisions.
Given that fact, any system of morality that requires adherents to have unquestioning faith in a necessarily imperfect belief system is fundamentally flawed.
Science makes no claim to be a system of morality. It is a system for determining "what is" rather than for determining "what should." People then ascribe meaning to that information according to their independent system of morality.
Religion very much claims to be a system of morality. And for most major religions, a frightening and dangerous one.
I do believe the level of discussion between the legal departments of the two companies would have been a bit more specific than the simplistic and vague statements that get posted around here.
The problem is that it usually goes something like this: We have several hundred patents which have some vague relation to what you're doing. Some of them you don't really infringe, others would probably be invalidated if challenged, but if you wanted to actually challenge them all you'd end up spending many millions of dollars. In the meantime the lawsuit makes the news, your stock price goes down and customers and developers shy away from your products. How about you just pay us the money instead?
Licensing one patent is patent licensing. Licensing dozens or hundreds of patents is a protection racket -- because the cost of actually analyzing and challenging them exceeds the cost of the protection money.
Clearly hardware is patent-able. So a novel arrangement of logic gates soldered onto a PCB should be patent-able, therefore the same configuration of logic gates implemented on a FPGA should be covered by that patent (it's the same logic gate arrangement after all)
You've already failed. Let's try it this way: Clearly hardware is patentable. So a novel arrangement of letters in movable type affixed to a printing press should be patentable, therefore the same story on a computer should be covered by that patent (it's the same text after all).
But obviously it isn't. You can't patent literature, in the same way as you ought not to patent software.
In the Nokia case Apple felt their IP was being undervalued and Nokia's overvalued. Cross-licensing is not automatically fair. Saying you hav to let us copy all your products if you want to use the 3G standard, also not fair.
But saying that you have to let us copy all your technologies if you want to use rounded corners or a touch screen is fair?
If Apple thinks the exchange is unfair then they're free to make products with rounded corners and no 3G and require Samsung to make products with 3G and no rounded corners and see which one succeeds in the marketplace, but you don't get to be a hypocrite and say that competitors have to license their stuff to us but we won't reciprocate.
What reason is there they will, altruism ?
Mutually assured destruction, obviously. If you sue someone, they sue you back, you both spend millions on litigation and end up worse off than if you had just negotiated a cross-license or reasonable royalties in the first place.
By contrast, Apple asked for this by firing the first shot.
Didn't Rambus win a while ago for doing pretty much the same thing?
Scenario: Your operative is in an unsecured location preparing for a mission. There is no SCIF in his vicinity. You learn new information which is relevant and must be communicated to him immediately.
It seems obvious that having a communications device which is as secure as practicable under those conditions is preferable to e.g. sending a completely unencrypted text message to his COTS cell phone.
Did you miss the "supermegacorp buried in bullshit" part of my theory?
I did not. The trouble is that that isn't what will happen. Nobody is really going to get a permanent injunction against the iPhone and then try to enforce it permanently. Especially not bidirectionally -- if you have an injunction against them and they get an injunction against you then you're both better off to sit down at the negotiating table. It's all just leverage in negotiating royalties or cross-licensing agreements. And it's largely a zero-sum game -- for every loser there is a winner who will be there lobbying Congress to keep their advantage, whether it's a megacorp crushing its smaller competitors or a patent troll putting a private tax on digital products and putting the money towards its lobbying efforts.
After all with their money they can crush the competition in a myriad of other ways besides patents and copyrights. Collusion and bribery for example.
Collusion doesn't work in a free market with low barriers to entry, which is what software would be without software patents. Bribery doesn't work without collusion because your competitors just pay bribes the other way and you don't get results. And both collusion and bribery are illegal.
Yeah, the NSA has a different security model than Apple.
For one thing, if the thing is really secure, it shouldn't matter that nefarious people get access to one -- that is one of the main things you need it to be secure against.
Of course, the way you do this is pretty obvious. You put plenty of memory in it but only read-only permanent storage which holds the OS and the device's unique private key, and store all other data "in the cloud" (i.e. on the NSA's secure server). You put a hardware AES engine on the CPU and have it encrypt everything in RAM. You have it establish an encrypted tunnel at all times to a secure building in spy central somewhere and send all other communications through that. Then you use two or three factor authentication to unlock the phone, which authenticates against the central server, and when the phone is locked the encryption key to decrypt most of memory is stored in the central location rather than on the phone. If the phone gets lost you disable its account on the server and it's instantly bricked because it can't even read its own memory, and it doesn't contain any sensitive data in permanent storage.
The ONLY way we are gonna get any kind of reforms after citizen united is if supermegacorp gets so buried in patent bullshit that they can't even release a single product here for having it tied up in court for a decade.
There is a major flaw in this strategy, which is that parasites are sticky.
Let's say we keep issuing ever more bullshit patents. Some coven of asshats the likes of Intellectual Ventures will acquire a few thousand of them and then extract several hundred million dollars each in settlements and lawsuits from shaking down the companies that actually make stuff. So then they've got a fat wallet and a proven business model -- what makes you think they aren't going to spend a billion dollars a year lobbying Congress to make sure it stays that way?
On top of that, you think the foreign manufacturers are going to put pressure on fixing the system, but it's completely the opposite. If the foreign companies are ignoring the patent trolls in their home countries, they can't sell the same products to the US market because they'll get sued the same as anybody else. So what happens? Market division. The US companies can't compete in foreign markets anymore because they have to abide the ridiculous US patent system, and the foreign companies stay out of the US because they don't want to have to either. Now ask yourself what incentive that drives for US companies -- if they fix the US patent system then foreign competition that had been avoiding the US in order to avoid litigation-induced bankruptcy will enter the market and compete with them. By contrast, if they can export the damage of software patents through treaties or foreign lobbying then they can use the warchests they originally built to balance one another to destroy their foreign competition.
Parasites are sticky and thrive on inefficiency. If you have fast-paced competition to be first to market, the parasites can't survive. If you let the market converge into a collection of oligopolies who stamp out innovative new competitors through litigation and resign themselves to paying the occasional tithe to a patent troll out of their monopoly rents, there could be nobody left without a vested interest in the status quo.
Didn't Handspring--now defunct--already do this like 12 years ago? There were all sorts of devices for their PDAS, including a GSM module.
When has that stopped anybody before? The lawyers will just try to find one little thing which is different and claim that it's new and patentable.
A decade ago we got a million patents that said something like "that thing everybody knows about, but on the internet." Now we're going to get a million patents that say something like "that thing everybody knows about, but on a mobile device." Because hey, adding that extra element lets you try to claim that nobody has ever done it before. And the patent office is profoundly lax in issuing rejections for obviousness.
This is a case in point: Making a computing device modular? Who could have ever thought of that? (People in the 1950's, that's who.) But you throw in "it's a phone too" and you've got yourself a patent application, apparently.
Regardless of the holding duration, that's a buyer.
Not if they already have the stock and are deciding when to sell it.
This is not a valid premise
The premise is not required, it is only a simplification to make the explanation easier. Let me put it a different way: Suppose that in the world without market makers, the majority of people would expect it to go up by approximately the specified amount. I am only trying to create a baseline in order to analyze the effects of modifying it.
Your further analysis basically comes to the conclusion that market makers remove capital from the system, depressing stock prices. Among other things, this ignores the set of participants who lose money on their trades.
Unless you are trying to argue that market makers on average have non-positive net profits, I'm not sure what this is supposed to be buying you.
Furthermore, anyone who doesn't want to pay for immediacy doesn't have to. You can submit a limit order for a specific price, and wait for the normal volatility of the market to come to your limit price.
That is something the seller can do regardless of the effect I'm talking about, so I'm not sure how it is relevant.
You are imagining the "penny shaving" from a buyer's perspective, but market makers view buying and selling as mostly symmetric (buy low, sell high; or sell high, buy low), hence why the spread is impacted and not the "price". In actuality, HF market makers are competing with each other for increasingly small penny-shavings, and they're able to do so because reduced latency leads to reduced risk.
I am imagining it from the perspective of anyone who is not a "market maker" -- someone who intends to hold the stock for more than a minute.
Let's take an example. A company has 1000 shares of stock. 800 are held by long-term investors who all intend to keep them for at least a year as long as nothing drastic happens. 200 are held by short term investors, who each intend to sell them within a short period of time X. The buyers for those 200 shares in turn intend to hold them each for another X period and then sell them again to another short-term investor, and so on indefinitely.
Now let's say that the "true" value of the company will increase by 2 cents/share every period, and everybody knows and expects that. Moreover, the short-term investors demand that profit of 2 cents per share every period at the current stock price, otherwise they'll take their money elsewhere and buy bonds or commodities or something. If there are no "market makers" then every period the short-term investors sell their stock for 2 cents/share more than they bought it for previously. Now enter the market makers. Let's say that for each trade, the market makers take an average of 1 cent/share.
Look at the chain it creates: The prospective buyers in the first period demand a 2c/share profit over the next period; they won't buy the shares for any more than what is necessary to achieve that. They know the market makers will take 1c, so what they demand is a 3c increase in the bid price over the period that they hold the shares. Based on the fundamentals the price will only go up 2c however. No one will pay more than the share is worth at the end of the period, so to create the required difference the immediate share price must go down by 1c.
But wait, there's more: The same dynamic will exist in the next period: The buyer in period two demands a 2c profit, which requires a 3c bid price increase over one period, and the natural value will again only increase 2c. In consequence the bid price for the stock in period two will be reduced by 1c/share. The buyer in period one realizes this, and still demanding a 2c profit, will now bid 2c/share less than due to the market makers -- 1c to cover the market maker's cut on this trade and 1c to cover the amount the next bidder demands to cover the market maker's cut on the next trade. And it goes on like this, driving down the current-day price of the stock because short-term investors are willing to pay less and less for it because the HFT market makers are over time predicted to consume that portion of their expected profits.
Regardless, a long-term investor is looking profit at least an order of magnitude above the cost of liquidity, so none of this should matter to Joe Sixpack and his 401k.
So you're looking at it from the perspective of the long-term investor and saying that it doesn't matter to them because they're going to buy for $N/share and sell for $2N/share, and the penny or two per share shaved off at the back by a market maker doesn't matter one lick to him because it's 0.5% of the profit. The trouble is that most of the buyers on the market at any given time are the short-term investors. Long-term investors go years without buying a share; they are not regularly available as buyers to other long-term investors. So when it comes time for a long-term investor to sell his shares, the value of the shares to the short-term investors is what matters. If the market makers materially decrease the value of the shares to those short term investors, it will correspondingly impact the price at which the long-term
It's possible that HFT is contributing to it to some small extent, but I would think that other factors would have more to do with it. For one thing, the super rich have highly diverse investments. They own stocks, but also government bonds, corporate bonds, land, commodities, etc. It is very rare that all of those things go down together, and in fact usually when one of them goes down another goes up. They also have a bunch of high-priced investment advisers whose entire job it is to make sure they aren't on the losing side of the deal and that if they might be then it's properly hedged.
On top of that, a lot of these people outright own a large private company, and a lot of those companies will regularly beat the market because they're run by the owners rather than officers overly focused on quarterly profits and afflicted with the principle-agent problem.
The thing that concerns me about HFT is that it's quite possibly a significant cause for why the stock market is doing so poorly. Think about it: If the HFTers shave a penny or so per share off of each trade between exchanges, future prospective buyers will be willing to pay that much less for their shares, because they know that when they ultimately sell them they'll suffer the same loss again to the HFTers. In consequence the seller is obligated to lower the price by that small pittance per share in order to make the trade, which sets the new market price for the stock. The value of the stock will continue to go down until it accounts for the amount of value the HFTers remove over the period of time that the average investor holds the stock.
The more the HFTers make over a given period of time, the less the stock is worth to "real" investors. More to the point, if the "true" value of the company does not increase faster than the speed at which HFTers remove value from stockholders through arbitrage, the value (and therefore stock price) of the company will continue slowly falling indefinitely, because no one will want to buy shares that will post-arbitrage be worth less than they paid for them. (In practice what will happen is that the value will fall until the only remaining investors holding the stock will be long-term investors, who reduce the value removed by HFTers because they no longer have a sufficient volume of trading to arbitrage, and so the stock will stabilize at a non-zero price where the value of the company is increasing faster than the arbitrage removes, but at a much lower value than it would be absent HFT.)
That isn't really much protection if you consider what drivers are. By their nature they have to have privileged access to hardware. So you don't have kernel mode yet, but you're a driver? You put some rootkit code in a noncritical register or buffer on whatever piece of hardware the driver is operating, then you instruct the device to write that data to a physical memory address where the kernel will eventually execute it in kernel mode.
And that's ignoring the "most" -- if some drivers are not virtualized then all you have to do is have the click-happy user install the rootkit purveyor's own unsigned driver, which is one of the sort that has to run in kernel mode.
But that's not caching at all, it's prefetch. The filesystem driver or the program itself can do that completely independent of the underlying cache architecture simply by reading the data in anticipation of it being needed soon.
And prefetch is no panacea. It makes huge mistakes. If you're playing that same game and there are 10GB worth of 300 different levels and it tries to load them into the cache when you're loading the first one, you're creating a bunch of unnecessary disk activity (and wearing out the SSD), and pushing things out of the cache that have actually been used recently, even though the user is very unlikely to want to use all 300 levels during one session.
Is it aware of the filesystem and keeps most accessed files, or just the most accessed blocks?
How is that an advantage? If I have e.g. a 5GB mail archive file which is my most accessed file, but I really only frequently access a few dozen MB of it which represent the most recently received mail, the performance will be much better if it will cache the frequently accessed blocks only and then use the rest of the SSD to cache other recently accessed blocks than if it uses up half the SSD to cache the whole mail archive even though I haven't accessed 7/8ths of it in the last two years.
People are having trouble understanding relativity. That is not an alternative.
Scenario 1: Apple continues selling iPhone 4 and does not release iPhone 5. Apple captures only 10% of the market, Android captures 60%, non-Android competitors capture 30%. If Android did not exist, Apple would capture 35% of the market and non-Android competitors would capture 65%.
Scenario 2: Apple brings out the iPhone 5. Apple captures 20% of the market, Android captures 55%, non-Android competitors capture 20%. If Android did not exist, Apple would capture 55% of the market and non-Android competitors would capture 45%.
In any scenario, the existence of additional competition reduces market share and/or margins that the other competitors would otherwise have.
The problem for Apple is that maintaining high margins by continually improving your product in a competitive market becomes a losing battle against diminishing returns. R&D effort into a new product line will sensibly go after the low-hanging fruit first, but once that is gone and all competitors have those features as a new baseline, it becomes more and more difficult as time passes to collect higher margins than competitors by releasing incremental iterations of existing products. The old products achieve most of the features that most people want, and the number of people willing to pay a $200 or $300 premium for version N+1 diminishes.
In order to maintain consistent profitability with tech products (as distinguished, in particular, from services), a company without an intrenched monopoly is required to continually enter or create new markets, as low-cost competitors naturally enter existing markets and drive down profits. Apple has had some success with this by getting into music players, then phones, now tablets. The question is, can they keep it up long-term? And the answer to that question is by no means clear.
Not really. Just because a signed kernel was loaded initially doesn't mean that's what is still running by the time the anti-malware program gets running. As soon as you have something malicious running with the ability to execute all privileged hardware instructions, there is no way to rely on anything anymore. The rootkit can just load a whole new kernel if it has to, or it can patch the running kernel in memory. The best anti-malware can do is to detect it before it executes in kernel mode and prevent it from executing. And trying to get the anti-malware running as early as possible only makes you more susceptible to rogue anti-malware.
1) Corporations would naturally have a BIOS password which only the IT department would know, and they would know better than to type it just because the dancing bunnies "for some reason" require you to install a boot loader.
2) If you can convince the user to reboot the computer, enter the BIOS and change the authorized boot loader by hand in order to see some lolcats, there is just nothing you can do to overcome that level of stupidity. You might as well worry about a denial of service attack whereby the website instructs the user to stick a fork in the electrical outlet. If the user is that stupid, you lose, and you deserve to lose.
Gates felt that Linux devs gained some sort of undue enrichment. It's a bogus and douchbag position, but it's not "rent seeking".
You're reading his bogus defense of his conduct and ignoring the conduct. Copyright and patent holders do this on a regular basis. They say that they created something and demand remuneration, but what they demand is far in excess of what they contributed. The patent holder patents the parts to his copy machine, but then tries to leverage the "legitimate" patent monopoly over replacement parts into a monopoly over copier service. The motion picture industry takes their oligopoly position in copyrighted motion pictures and tries to leverage it into control over the distribution channels, and then over all consumer electronics.
If Microsoft breaks ACPI for Linux, they break more than they built. They are (to use the Microsoft camp's philosophy and terminology) "stealing" the benefit to the hardware makers who developed ACPI of selling ACPI-functional hardware to users of non-Microsoft operating systems. They are also ignoring the legitimate work done by the Linux camp to make ACPI work on Linux, as though somehow only Microsoft's efforts to make ACPI work for Windows are legitimate and require compensation and consideration but any efforts by a third party to make it work for another operating system can be ignored in the calculations.