Yes, it is a common mistake. That's why Apple has told people not to do what you're describing at pretty much every WWDC networking talk for as far back as I can remember, and devoted an entire chapter in their networking documentation to that subject.
I'm not seeing anything like the above in the Android documentation, which may explain why this is a much more common problem on that platform. And pretty much all the sample code I see on places like Stack Overflow and GitHub are wrong, which further compounds the problem. Want this problem to go away? Go to all those Stack Overflow pages and GitHub pages and flag them as inappropriate. Then convince someone to document how to do it the *right* way.
That's true only if you have previously explicitly agreed to the company's choice of venue/forum in a contract. Otherwise, as a rule, IIRC, you can sue in any of A. the court nearest the plaintiff, B. the court nearest the defendant, or C. the court nearest where the injury occurred. And even when the company has a choice-of-forum clause in a contract, the judge has the right to declare it unenforceable if it would pose undue hardship to an individual plaintiff suing a large, multinational corporation or would allow such a corporation to violate the public policy of the state where the plaintiff lives because of substantively weaker laws in the defendant's state (Doe 1 v. AOL LLC.
That's nothing. I bought a bunch of microphones from a Chinese manufacturer a couple of years ago. The power supply circuit boards were machine-built, AFAIK, so they were consistent. All that they had to do by hand was screw a single circuit board down with... maybe four screws, screw the case together with... I think two screws, and shove on a knob or two on the front. They might have had to snap a power jack into the back or something. We're talking dead simple here.
I'll let you guess what percentage of them had at least one screw rolling around in the box. Hint: it was not a single-digit percentage. Most companies can't cope with that sort of return rate.
If you want quality out of China, you'll only get it if you have enough volume that they will care if they lose your business. And you will have to do random inspections of their factories, which means having employees on the ground in China. Fail to meet either of those requirements and, assuming what I've seen is typical, it would probably be cheaper for them to ship you the parts and for you to pay to have them assembled in the U.S. once you factor in the astounding return rate. Assuming that you do proper QA in the U.S., you'll be disassembling and reassembling a quarter of them anyway, so you might as well eliminate the redundant assembly step and just assemble it over here. Cheaper import duties, that way, too.
You can still get kernel sources. Here's the kernel for 10.8.2. AFAIK, OS X's xnu kernel has never been closed source unless you count the delay after each major OS release while they clean stuff up....
And by that, of course, you mean that they ship most of them by air during the initial ramp-up period while they're still filling up the channel with stock.
If they're like most companies, the first few weeks involve a lot of air shipments because it takes about a month (or a little bit over) to get products to the U.S. from China by boat. After that, assuming you guessed the average demand correctly, you should be able to keep the channel basically filled with weekly or monthly boat deliveries, and limit the use of air shipments to dealing with any unexpected surges in popularity.
The difference between that electricity and a light bulb is that at low power bulb can light a room with much less power than an old style 60W bulb.
Except that it can't. Low-power bulbs (at least the ones that I consider to use "much less power") necessarily have massive peaks and troughs in their spectrum. If you've ever taken photographs under fluorescent light or gone clothes shopping only to find that the shirt that was blue in the store is green outside, you'll realize just how unforgiving high-efficiency lighting is.
With halogen bulbs, you can light a room with a bit less power than a traditional incandescent (about 28% less wattage per lumen), but not a lot less, at least by comparison with the near-order-of-magnitude improvements you get from LED lighting and fluorescent bulbs. It is not currently possible to use a lot less power (read "whole-number factors") without severely compromising color accuracy.
More to the point, we might be able to come up with some clever new filament designs that result in another small win, but we're probably pretty close to the limit. The inefficiency comes from producing infrared and, to a limited degree, ultraviolet light. You can't eliminate that without either chopping off the ends of your light spectrum (which means color inaccuracy near both ends) or moving to something that produces artificial peaks that all fall within the range of human vision (like phosphor-based incandescent and LED bulbs that are inaccurate throughout the spectrum). That means they probably cannot restrict things very much further without outright banning all full-spectrum lighting.
Or the journalist doesn't bother to look at the product roadmap to realize that the vendors are already working towards the standard, and by the time it actually gets implemented, the vendors' products are expected to be compliant.
You misunderstood the summary (which is not entirely your fault—it was pretty hard to read up until the last paragraph, which explained it reasonably well). They didn't limit bandwidth. They increased the minimum bandwidth that a card has to achieve if it wants to draw a certain amount of power.
The regulations divide graphics cards up by their peak bandwidth (or maybe average bandwidth—I'm not sure which). Low-bandwidth devices are not allowed to consume much power. The next tier of devices have higher peak bandwidth and are allowed to consume more power. And so on. They increased the minimum bandwidth requirement for the highest-tier category (the category containing the fastest cards). The result is that video cards that previously fell into the top-bandwidth bucket now fall into a lower-bandwidth bucket and are no longer allowed to draw as much power.
So the card vendors' options are: A. find a way to draw less power or B. increase the peak bandwidth so that they qualify as a high-end graphics card again. And what the article is saying is that none of the current or upcoming high-end cards have enough bandwidth to fall into that top category, but they all draw too much power to meet the criteria for the next bucket down.
Very few got the internet, or the prevalence of pocket computing and connectivity that we take for granted 20 years later.
Star Trek had the basic concept of portable computing in the late 1960s, albeit crudely. And I'm pretty sure that there were folks predicting it long before that.
Mark Twain predicted the Internet in the late 1800s. Not precisely, of course—who would have thought that text-based communications would actually make a comeback—but he pretty much described the concept of a worldwide communication network with webcams where you could see and hear what was going on around the world... in an era when computers were mechanical devices, when television was basically still in the hypothetical stage, etc.
What people didn't predict was that we would clog up those pipes with advertising....
Unfortunately, members of the parties have approximately zero chance of being who they claim to be during the election. As soon as there's a vote, they vote for or against whatever their party says to support or block, even if it is the opposite of what they said just a few weeks before. And because the party won't let anyone run against them during the primaries, once you get a party member in office, you're never going to get rid of him/her unless you can miraculously convince enough people to vote in someone from the opposite party which, thanks to gerrymandering, is almost impossible.
The same is said of consensual sex with a minor: anyone under legal age is incapable of consenting. A 22 that has sex with a girl 17 years old, 364 days at 10PM is a felon who must register for the rest of his life as sex offender, but if they go to a movie first and then get it on at 12:01AM he's in the clear. (I'll leave it to other commenters to come up with a snarky comment)
Girl: If I'm gonna do that, the least you can do is take me to a movie first.
Guy: At your age, I'd be crazy not to.
But seriously, the problem with the boundary conditions is that you have a choice: set boundary conditions that are too lax and some people will get away with being dirtbags; set boundary conditions that are too loose and some people will get jailed for no good reason; set boundary conditions in the middle, and both of the above will happen.
The better solution is to have different laws depending on the situation. For example, incest, abuse of a minor in your care, etc. are separately crimes when they involve someone under 18, period. This means that for those situations, you don't need the statutory rape laws; they're redundant. So if you beef up the law by making other always-abusive situations illegal when it involves anyone 18-and-under, the statutory rape laws become less important, and it won't hurt to weaken them so that the only absolute bans are on sex that is way over the line of acceptable behavior, i.e. lowering the minimum age and allowing moderately wide age gaps.
Alternatively, change the law to ban prosecution without the consent of the aggrieved minor, and make evidence of any pressure on said minor by the authorities be grounds for dismissal of the charges. And give the aggrieved minor the right to accept or reject any proposed sentencing. In other words, change it to a "no harm, no foul" law—de minimis non curat lex and all that.
The remarkable thing is that in the U.S., the same thing happens, and we've been blaming it on the two-party system. Apparently, such blatant disregard for voters is not caused by the number of parties, but by the mere existence of parties.
Actually, some things that are ordinarily fallacies cease to even be fallacies in the context of a political debate. For example, ad hominem attacks are not inherently fallacies in the context of a political debate because the desired outcome of the debate is not to decide whose stated position is right, but rather who would be the better choice for that office.
A classic example of a non-fallacious fallacy in political debates is the appeal to hypocrisy. Such an appeal is fallacious when used to evaluate the validity of the candidate's position. However, the appeal is not entirely fallacious with regard to the debate as a whole because what actually matters is the way the candidate will likely actually vote, not the way the candidate says he or she will vote.
In fact, to the degree that a significant number of appeals to hypocrisy can be made against a politician, it usually dooms the candidate in question, and for good reason. If you don't really know where the candidate stands—if he or she says one thing and does another—that person is a really bad choice for any office.
If you loan them your car then you're an accessory to the crime.
No, you aren't, or at least not in the U.S. You are only an accessory if it can be shown that you had actual knowledge that the person who borrowed your car intended to rob a bank. If someone asks to borrow my car and then, without my knowledge, uses it to rob a bank, I am not an accessory.
Similarly, you are only an accessory if you knew that loaning him or her your car would help him or her in committing that crime or evading capture in some way. If somebody tells me he or she is thinking about robbing a bank and then, in a separate conversation, asks to borrow my car to go get milk, I am not an accessory even if he or she then robs the bank using my car.
You have to have not only knowledge of the crime, but also intent to aid in the commission of the crime.
Even those $300 non-PC computers would be about $639 in today's dollars, which is still more than six times what the cheapest non-Windows netbooks cost today. And they weren't nearly as capable as PCs. Most of the cost savings (except for the CPU) came from ingenious hacks that reduced the cost of the hardware at the expense of maintainability and the ability to upgrade the design later. For example:
Commodore used software emulation instead of a real UART, which limited them to about 2400 baud and made them particularly challenging if the baud rate on the remote device was off just a bit.
The Apple II's NTSC output wasn't actually NTSC compliant (read "can't be safely recorded or broadcast") and remained so through at least the IIe. When they added PAL support, it required different graphics hardware.
Now you can certainly argue that those things didn't matter, and for most people, you'd be right, but the same argument that says that the $300 computer was equal to the $2,000 computer in the 1980s applies to touchscreen tablets versus desktop computers today.
Also, the only reason prices aren't $100k per CPU today is because of economies of scale. Today, one person could design a basic Verilog model of a 6502-compatible chip in a matter of days or single-digit weeks, and even at the time, it probably took double-digit engineer years. Today, an Intel chip takes engineer-millennia. The R&D involved is orders of magnitude greater, and therefore, the number of chips they have to sell just to break even is also much, much greater.
Proper operating systems spin up fans and, if necessary, throttle the CPU to keep things well within a safe range. Any computer that crashes under load was probably assembled wrong.
Laptops are simply not reliable and not repairable or modifiable by most people....
Maybe you should try spending more than $300 on a laptop next time. I just replaced my previous laptop after more than five years of reliable service, except for one dead hard drive, which was trivial to replace. And the only reason I replaced the machine at all is that it is an anachronism that won't run some modern software....
Kill, no. Mortally wound, perhaps. Think about it this way: right now, you can get cheap PCs for a few hundred bucks. Adjusted for inflation, computers in the mid-1980s ranged from about $3000-$6000 in today's dollars. Now think back to high school economics class and remember the discussion of economies of scale, then think about how few parts from modern tablets are actually used in a typical desktop computer.
Oh, come on. HTML5 offline manifests and local storage work positively flawlessly compared to contentEditable support, undo management, copy-and-paste handling, DOM Ranges....
These days, a good day of web app development is one in which I discover fewer than one critical browser bug every two or three hours of coding. I won't tell you what a bad day looks like because I don't want to crush anyone's soul....
"The fact that there are no disc-shaped aircraft in the skies today, though, suggests that the USAF's flying saucer efforts probably never got past the prototype stage."
or they work so good that only blurry and shaky videos exist of them flying around and terrorizing cows
Slashdot Mirror
Yes, it is a common mistake. That's why Apple has told people not to do what you're describing at pretty much every WWDC networking talk for as far back as I can remember, and devoted an entire chapter in their networking documentation to that subject.
I'm not seeing anything like the above in the Android documentation, which may explain why this is a much more common problem on that platform. And pretty much all the sample code I see on places like Stack Overflow and GitHub are wrong, which further compounds the problem. Want this problem to go away? Go to all those Stack Overflow pages and GitHub pages and flag them as inappropriate. Then convince someone to document how to do it the *right* way.
I'm pretty sure there's a joke in there somewhere.
That's true only if you have previously explicitly agreed to the company's choice of venue/forum in a contract. Otherwise, as a rule, IIRC, you can sue in any of A. the court nearest the plaintiff, B. the court nearest the defendant, or C. the court nearest where the injury occurred. And even when the company has a choice-of-forum clause in a contract, the judge has the right to declare it unenforceable if it would pose undue hardship to an individual plaintiff suing a large, multinational corporation or would allow such a corporation to violate the public policy of the state where the plaintiff lives because of substantively weaker laws in the defendant's state (Doe 1 v. AOL LLC.
That's nothing. I bought a bunch of microphones from a Chinese manufacturer a couple of years ago. The power supply circuit boards were machine-built, AFAIK, so they were consistent. All that they had to do by hand was screw a single circuit board down with... maybe four screws, screw the case together with... I think two screws, and shove on a knob or two on the front. They might have had to snap a power jack into the back or something. We're talking dead simple here.
I'll let you guess what percentage of them had at least one screw rolling around in the box. Hint: it was not a single-digit percentage. Most companies can't cope with that sort of return rate.
If you want quality out of China, you'll only get it if you have enough volume that they will care if they lose your business. And you will have to do random inspections of their factories, which means having employees on the ground in China. Fail to meet either of those requirements and, assuming what I've seen is typical, it would probably be cheaper for them to ship you the parts and for you to pay to have them assembled in the U.S. once you factor in the astounding return rate. Assuming that you do proper QA in the U.S., you'll be disassembling and reassembling a quarter of them anyway, so you might as well eliminate the redundant assembly step and just assemble it over here. Cheaper import duties, that way, too.
You can still get kernel sources. Here's the kernel for 10.8.2. AFAIK, OS X's xnu kernel has never been closed source unless you count the delay after each major OS release while they clean stuff up....
And by that, of course, you mean that they ship most of them by air during the initial ramp-up period while they're still filling up the channel with stock.
If they're like most companies, the first few weeks involve a lot of air shipments because it takes about a month (or a little bit over) to get products to the U.S. from China by boat. After that, assuming you guessed the average demand correctly, you should be able to keep the channel basically filled with weekly or monthly boat deliveries, and limit the use of air shipments to dealing with any unexpected surges in popularity.
I accidentally did that in a test drive once, too. Needless to say, I did not buy that particular model of vehicle....
Except that it can't. Low-power bulbs (at least the ones that I consider to use "much less power") necessarily have massive peaks and troughs in their spectrum. If you've ever taken photographs under fluorescent light or gone clothes shopping only to find that the shirt that was blue in the store is green outside, you'll realize just how unforgiving high-efficiency lighting is.
With halogen bulbs, you can light a room with a bit less power than a traditional incandescent (about 28% less wattage per lumen), but not a lot less, at least by comparison with the near-order-of-magnitude improvements you get from LED lighting and fluorescent bulbs. It is not currently possible to use a lot less power (read "whole-number factors") without severely compromising color accuracy.
More to the point, we might be able to come up with some clever new filament designs that result in another small win, but we're probably pretty close to the limit. The inefficiency comes from producing infrared and, to a limited degree, ultraviolet light. You can't eliminate that without either chopping off the ends of your light spectrum (which means color inaccuracy near both ends) or moving to something that produces artificial peaks that all fall within the range of human vision (like phosphor-based incandescent and LED bulbs that are inaccurate throughout the spectrum). That means they probably cannot restrict things very much further without outright banning all full-spectrum lighting.
Or the journalist doesn't bother to look at the product roadmap to realize that the vendors are already working towards the standard, and by the time it actually gets implemented, the vendors' products are expected to be compliant.
You misunderstood the summary (which is not entirely your fault—it was pretty hard to read up until the last paragraph, which explained it reasonably well). They didn't limit bandwidth. They increased the minimum bandwidth that a card has to achieve if it wants to draw a certain amount of power.
The regulations divide graphics cards up by their peak bandwidth (or maybe average bandwidth—I'm not sure which). Low-bandwidth devices are not allowed to consume much power. The next tier of devices have higher peak bandwidth and are allowed to consume more power. And so on. They increased the minimum bandwidth requirement for the highest-tier category (the category containing the fastest cards). The result is that video cards that previously fell into the top-bandwidth bucket now fall into a lower-bandwidth bucket and are no longer allowed to draw as much power.
So the card vendors' options are: A. find a way to draw less power or B. increase the peak bandwidth so that they qualify as a high-end graphics card again. And what the article is saying is that none of the current or upcoming high-end cards have enough bandwidth to fall into that top category, but they all draw too much power to meet the criteria for the next bucket down.
Star Trek had the basic concept of portable computing in the late 1960s, albeit crudely. And I'm pretty sure that there were folks predicting it long before that.
Mark Twain predicted the Internet in the late 1800s. Not precisely, of course—who would have thought that text-based communications would actually make a comeback—but he pretty much described the concept of a worldwide communication network with webcams where you could see and hear what was going on around the world... in an era when computers were mechanical devices, when television was basically still in the hypothetical stage, etc.
What people didn't predict was that we would clog up those pipes with advertising....
Blasphemer! This new concept of zero is a heresy against Ra!
Too soon?
Unfortunately, members of the parties have approximately zero chance of being who they claim to be during the election. As soon as there's a vote, they vote for or against whatever their party says to support or block, even if it is the opposite of what they said just a few weeks before. And because the party won't let anyone run against them during the primaries, once you get a party member in office, you're never going to get rid of him/her unless you can miraculously convince enough people to vote in someone from the opposite party which, thanks to gerrymandering, is almost impossible.
Girl: If I'm gonna do that, the least you can do is take me to a movie first.
Guy: At your age, I'd be crazy not to.
But seriously, the problem with the boundary conditions is that you have a choice: set boundary conditions that are too lax and some people will get away with being dirtbags; set boundary conditions that are too loose and some people will get jailed for no good reason; set boundary conditions in the middle, and both of the above will happen.
The better solution is to have different laws depending on the situation. For example, incest, abuse of a minor in your care, etc. are separately crimes when they involve someone under 18, period. This means that for those situations, you don't need the statutory rape laws; they're redundant. So if you beef up the law by making other always-abusive situations illegal when it involves anyone 18-and-under, the statutory rape laws become less important, and it won't hurt to weaken them so that the only absolute bans are on sex that is way over the line of acceptable behavior, i.e. lowering the minimum age and allowing moderately wide age gaps.
Alternatively, change the law to ban prosecution without the consent of the aggrieved minor, and make evidence of any pressure on said minor by the authorities be grounds for dismissal of the charges. And give the aggrieved minor the right to accept or reject any proposed sentencing. In other words, change it to a "no harm, no foul" law—de minimis non curat lex and all that.
The remarkable thing is that in the U.S., the same thing happens, and we've been blaming it on the two-party system. Apparently, such blatant disregard for voters is not caused by the number of parties, but by the mere existence of parties.
Actually, some things that are ordinarily fallacies cease to even be fallacies in the context of a political debate. For example, ad hominem attacks are not inherently fallacies in the context of a political debate because the desired outcome of the debate is not to decide whose stated position is right, but rather who would be the better choice for that office.
A classic example of a non-fallacious fallacy in political debates is the appeal to hypocrisy. Such an appeal is fallacious when used to evaluate the validity of the candidate's position. However, the appeal is not entirely fallacious with regard to the debate as a whole because what actually matters is the way the candidate will likely actually vote, not the way the candidate says he or she will vote.
In fact, to the degree that a significant number of appeals to hypocrisy can be made against a politician, it usually dooms the candidate in question, and for good reason. If you don't really know where the candidate stands—if he or she says one thing and does another—that person is a really bad choice for any office.
...then it will just reappear, possibly a piece at a time, attached to some appropriations bill for homeless battered women's shelters.
No, you aren't, or at least not in the U.S. You are only an accessory if it can be shown that you had actual knowledge that the person who borrowed your car intended to rob a bank. If someone asks to borrow my car and then, without my knowledge, uses it to rob a bank, I am not an accessory.
Similarly, you are only an accessory if you knew that loaning him or her your car would help him or her in committing that crime or evading capture in some way. If somebody tells me he or she is thinking about robbing a bank and then, in a separate conversation, asks to borrow my car to go get milk, I am not an accessory even if he or she then robs the bank using my car.
You have to have not only knowledge of the crime, but also intent to aid in the commission of the crime.
Even those $300 non-PC computers would be about $639 in today's dollars, which is still more than six times what the cheapest non-Windows netbooks cost today. And they weren't nearly as capable as PCs. Most of the cost savings (except for the CPU) came from ingenious hacks that reduced the cost of the hardware at the expense of maintainability and the ability to upgrade the design later. For example:
Now you can certainly argue that those things didn't matter, and for most people, you'd be right, but the same argument that says that the $300 computer was equal to the $2,000 computer in the 1980s applies to touchscreen tablets versus desktop computers today.
Also, the only reason prices aren't $100k per CPU today is because of economies of scale. Today, one person could design a basic Verilog model of a 6502-compatible chip in a matter of days or single-digit weeks, and even at the time, it probably took double-digit engineer years. Today, an Intel chip takes engineer-millennia. The R&D involved is orders of magnitude greater, and therefore, the number of chips they have to sell just to break even is also much, much greater.
Proper operating systems spin up fans and, if necessary, throttle the CPU to keep things well within a safe range. Any computer that crashes under load was probably assembled wrong.
Maybe you should try spending more than $300 on a laptop next time. I just replaced my previous laptop after more than five years of reliable service, except for one dead hard drive, which was trivial to replace. And the only reason I replaced the machine at all is that it is an anachronism that won't run some modern software....
Kill, no. Mortally wound, perhaps. Think about it this way: right now, you can get cheap PCs for a few hundred bucks. Adjusted for inflation, computers in the mid-1980s ranged from about $3000-$6000 in today's dollars. Now think back to high school economics class and remember the discussion of economies of scale, then think about how few parts from modern tablets are actually used in a typical desktop computer.
Wrong hole.
Oh, come on. HTML5 offline manifests and local storage work positively flawlessly compared to contentEditable support, undo management, copy-and-paste handling, DOM Ranges....
These days, a good day of web app development is one in which I discover fewer than one critical browser bug every two or three hours of coding. I won't tell you what a bad day looks like because I don't want to crush anyone's soul....
They all crashed in Roswell....