Translation: audio and video devices over USB don't work properly. (And many of them do; AFAIK, the problems are specific to USB 2.0 devices, and probably only in the presence of USB 1.1 devices like keyboards.)
Compared with where things were a few months ago, that's a near-miraculous state. The prior state could best be described as "I plugged in a USB 1.x device through a USB 2.0 powered hub, and Ethernet stopped working until I unplugged the keyboard.":-)
As I understand it, many of the problems stem from serious interrupt latency bugs in other parts of the kernel (e.g. the SD card driver). Until they litter the Broadcom stack with preemption points, you're going to have some of these problems, and possibly even then. Many of the other problems are caused by a horrible USB driver, which I think they're in the process of rewriting. But for non-A/V use, the Pi's USB works quite well as far as I've seen, at least with the latest bits. The split transaction support is dramatically improved, for example, so you almost never see the stuck key problems that plagued the Pi just a few months back.
A darknet relies on being invisible to the public, but not on anonymous towards each other : in fact, to avoid discovery, it's important to know whom you are sharing with, so that you are not accidentally sharing with a MAFIAA spy.
I would argue that this is not a real darknet. A proper darknet is fully anonymous, even from other people on the darknet. More precisely, a darknet does not hide your darknet identity from others on the darknet, but does prevent others from using that identity to gain your real-world identity.
What you're describing is just a highfalutin word for basic encryption.
Everything, even ethernet, is USB connected, and the USB controller is broken.
There were some major problems when mixing certain low-speed (10 Mbps) devices with high-speed devices, but that problem was basically fixed with a device firmware update and maybe a kernel update. I haven't had any problems since telling my Pi to update itself.
That said, last I checked, the version of Raspbian that you can actually download in image form was way out of date and did not include those fixes. So you do have to explicitly tell it to update its firmware bits right out of the box.
DNA is highly circumstantial evidence by itself. In the absence of means and a motive, it should not be enough evidence to arrest anyone, much less convict someone.
Unless, of course, that DNA was found in meiotic form inside a body cavity, in which case it is much less circumstantial.
Eventually, some of the rare earth metals used in electronics manufacturing may be depleted to the point where this would make sense to avoid a technological event horizon of sorts. At that point, there would be advantages to also bringing back large quantities of gold, copper, lithium, and possibly aluminum in the same shipment (since the overhead for transporting stuff from Mars is huge, and the cost of leaving empty space in the cargo hold is equally huge).
Actually, that's not the most likely explanation. The most likely explanation is that the DNA sample that they believed came from the assailant just happened to be at the crime scene for any of a million other reasons.
The second most likely explanation is that there's too much homogeneity in the DNA of the Dutch people—too pure a gene pool. DNA tests are likely to have a much higher false positive rate there than in a less homogeneous population like the U.S., and for a small number of loci, that could result in an unacceptable false match rate.
Contamination of the sample after it was taken is way down around number twenty or so in terms of likelihood.
Obama won the popular vote by a significant margin ( 3,476,775 votes)...
Only in the most pedantic sense. If the electorate had contained only a hundred people, that would have been about a 2.6 vote margin. A truly significant margin, in my mind, is more like 60/40. Also, nearly half of the eligible voters didn't even bother to show up.
Basically, what this election showed us is that A. many people could not tell much difference between the two candidates, so they voted to keep the guy with a four-year track record as President instead of voting in the guy with no real record in national politics, and B. among everyone who saw a difference between the two candidates, there were about as many people who liked/hated each candidate.
In other words, both candidates were remarkably bad at appealing to the masses.
They got a letter from me pointing out a few gaping flaws. The biggest ones were:
Starting a mere twelve-year period when a work is first created means that larger literary works written by individuals in their spare
time would, without renewal, be partially out of copyright protection before they are completed. The flexibility for unpublished papers would need to be a fair use exception, not a duration thing. Twenty-eight years was long enough pre-1976 only because the clock didn't start until the first publication date.
The periods are too complex—12, 12, 6, 6, 10. That's not different enough from 12, 12, 12, 12 to be worth the complexity.
The complexity of determining revenue would be too much hassle for individuals and would result in tricks to hide the revenue by companies.
I concluded by saying that the purpose of reform should be ensuring that works aren't protected once they're no longer making money. In my mind, there's no benefit to mandating that every possible work is out of copyright in a given maximum number of years. If Disney or Lucas wants to pay a fortune to keep exclusive rights to Steamboat Willie, that's not really a problem so long as 99% of works go into the public domain in a timely manner.
In other words, it is a good start, but there are enough rough edges that the paper needs more review before it serves as any sort of starting point for legislation.
You're confusing the right of free speech with the right to be heard. The former is protected, the latter is not. E.g., you have the right to say something. You don't have the right to force me to hear you.
No, you're misunderstanding me. I don't have the right to be heard in that you have the right to not listen. I do have the right to be hearable, however. That's a crucial distinction.
For example, it's okay for a politician to buy up a bunch of first-adjacent or last-adjacent ad slots to air their attack ads. It's not okay for them to buy up every adjacent ad slot on every station and fill them all with pictures of poodles so that his or her opponent cannot get any air time at all.
To a point. Beyond a certain point, however, an organization of people can amass the ability to speak at sufficient volume (both in loudness and in quantity) to effectively drown out dissenting voices. Thus, in order to guarantee free speech for the individual, to some degree, the speech of large groups must be kept in check.
This applies whether the large group is the government, a PAC, a corporation, a union, or pretty much any other group of people who have united for a single purpose. For example, the U.S. government is structured very deliberately to minimize tyranny of the majority.
Similarly, commercial speech (e.g. advertising) has limits that require a degree of truthfulness. Commercial speech is not really different from any other speech, apart from the fact that the people doing the speaking have the presumption of authority and the volume to back up that presumption. We just have more laws to cover that corner case because it is fairly common and is often egregious.:-)
No, that's basically how FaceTime works, too. You call someone on the phone, and if the other end is an iOS device, there's a button you can push to switch the call over to FaceTime. It might require an iCloud account (I'm honestly not sure), but they're free, so that's not particularly important....
Have you looked into the wireless broadband carriers? I mean the ones with an antenna on the roof, not cellular. They cover most of the greater Silicon Valley area unless you're up in the mountains or something.
You do realize that you can choose the packaging, right? You can't usually choose whether to use PPPoE or PPPoA, for the most part. It's like your only shipper is a really greedy pack-and-ship firm that deliberately uses oversized boxes even for non-fragile items that are already in their own boxes.
I'm pretty sure you'll find that any OS less than a decade old supports path MTU discovery for IPv4. So that should only be an issue if you're running some really ancient OS.
Actually, he was only about half right. Used tastefully and in moderation, dissonance can create mood in ways that consonance cannot easily match. As with nearly all of the musical techniques that he argued were historically dissonant (with the exception of basic polyphony), however, used in excess, it sounds like crap.
IMO, the key to the tasteful use of dissonance is to make sure that the dissonance is not the focus. On the one extreme, you might have the subtle use of dissonant suspension and release in secondary parts of a complex orchestral work to set the mood. On the other extreme, you might have a highly dissonant piece of music used as the background sound behind a Civil War battle. In both cases, the listener is focused on something else, whether that something else is a traditional melodic line or a bunch of people shooting each other in a horrible, bloody battle.
Incidentally, most folks (statistically) don't like heavy metal, highly dissonant jazz, bebop, etc. even to this day. Those genres and subgenres all serve a useful purpose when it comes to expanding the musical universe, and over time, those experimental ideas will get incorporated into more mainstream music in much more subtle and toned-down ways, but that doesn't mean that most people will ever find the experimental music itself enjoyable to listen to.
No, these days, I'm pretty sure the only way to survive getting a PhD in anything with your sanity intact involves a copious amount of the strong stuff....
Your efficiency of not doing work is like measuring the MPG you get idling in your driveway.
Not at all. You idle in your driveway and at traffic lights for a couple of minutes per trip. An average processor while running a typical word processor or web browser is idle up to 98% of the time while waiting for the user to click or type something. Even if you drive on highways in bumper-to-bumper traffic every day, your engine's idle percentage isn't going to be that bad. And even when it is relatively non-idle, it usually isn't using more than 20-30% of a single CPU core unless you're playing a game, playing video, or doing some other similarly CPU-intensive task.
Measuring the power consumption with a benchmark is like measuring fuel economy at 120 MPH uphill all the way while towing a trailer. It isn't real-world use. When you measure a car's fuel economy, you measure it part of the time at low speed, part of the time at high speed, part of the time stoppping and starting, and part of the time idling, because that's how cars are actually driven. But even that isn't the way your computer is used.
Measuring CPU power usage correctly is more like measuring the fuel economy impact of a car's air conditioner. You don't measure it in 150 degree heat so that the air conditioner compressor runs 100% of the time, because that's not the way it is used in the real world. In the real world, the compressor runs for one minute and then disengages for ten.
And computers have always needed to be more efficient when idle. The only reason any of this stuff is happening is because Intel is starting to panic at the possibility of their share of the computing market drying up because of their idle wattage. Just look at Dell prototyping 64-bit ARM-based servers, and you'll understand why Intel is running scared.
If anything, idle power consumption matters even more for beefy hardware than for things like cell phones. You can always put a bigger battery in a cell phone and nobody is going to scream. When you're dealing with server farms that consume millions of dollars per year in electricity, significant reductions in idle power consumption can save a lot of money even if it means adding more hardware (so long as that extra hardware doesn't bring the idle power consumption back up to where it was before, that is). Granted, demand-based cluster scaling can help with power consumption, but those efficiency wins come at a cost in terms of the ability to rapidly respond to large changes in demand.
Haswell, assuming it lives up to the hype, should be a big step in the right direction, but since we're talking about hardware that (from what I've read, anyway) won't ship for several more months, and for which no public benchmark data exists AFAIK, it's really too soon to say for sure.
That's a false comparison, though. If users mostly ran benchmarks 24x7, that would be a good test of efficiency. The reality, however, is that CPUs mostly sit idle, so to compute average efficiency, you have to factor that in.
Granted, a faster CPU that can reach an idle state sooner can be more efficient than a slower CPU that runs at full bore for a longer period of time, but only if the idle wattage is fairly similar.
Slashdot Mirror
Translation: audio and video devices over USB don't work properly. (And many of them do; AFAIK, the problems are specific to USB 2.0 devices, and probably only in the presence of USB 1.1 devices like keyboards.)
Compared with where things were a few months ago, that's a near-miraculous state. The prior state could best be described as "I plugged in a USB 1.x device through a USB 2.0 powered hub, and Ethernet stopped working until I unplugged the keyboard." :-)
As I understand it, many of the problems stem from serious interrupt latency bugs in other parts of the kernel (e.g. the SD card driver). Until they litter the Broadcom stack with preemption points, you're going to have some of these problems, and possibly even then. Many of the other problems are caused by a horrible USB driver, which I think they're in the process of rewriting. But for non-A/V use, the Pi's USB works quite well as far as I've seen, at least with the latest bits. The split transaction support is dramatically improved, for example, so you almost never see the stuck key problems that plagued the Pi just a few months back.
I would argue that this is not a real darknet. A proper darknet is fully anonymous, even from other people on the darknet. More precisely, a darknet does not hide your darknet identity from others on the darknet, but does prevent others from using that identity to gain your real-world identity.
What you're describing is just a highfalutin word for basic encryption.
There were some major problems when mixing certain low-speed (10 Mbps) devices with high-speed devices, but that problem was basically fixed with a device firmware update and maybe a kernel update. I haven't had any problems since telling my Pi to update itself.
That said, last I checked, the version of Raspbian that you can actually download in image form was way out of date and did not include those fixes. So you do have to explicitly tell it to update its firmware bits right out of the box.
Or she could have sat in the same chair.
DNA is highly circumstantial evidence by itself. In the absence of means and a motive, it should not be enough evidence to arrest anyone, much less convict someone.
Unless, of course, that DNA was found in meiotic form inside a body cavity, in which case it is much less circumstantial.
Is it a black hole?
Eventually, some of the rare earth metals used in electronics manufacturing may be depleted to the point where this would make sense to avoid a technological event horizon of sorts. At that point, there would be advantages to also bringing back large quantities of gold, copper, lithium, and possibly aluminum in the same shipment (since the overhead for transporting stuff from Mars is huge, and the cost of leaving empty space in the cargo hold is equally huge).
And then Hillary said, "Good plan. You bring the rocket. I'll bring Santorum."
Actually, that's not the most likely explanation. The most likely explanation is that the DNA sample that they believed came from the assailant just happened to be at the crime scene for any of a million other reasons.
The second most likely explanation is that there's too much homogeneity in the DNA of the Dutch people—too pure a gene pool. DNA tests are likely to have a much higher false positive rate there than in a less homogeneous population like the U.S., and for a small number of loci, that could result in an unacceptable false match rate.
Contamination of the sample after it was taken is way down around number twenty or so in terms of likelihood.
Only in the most pedantic sense. If the electorate had contained only a hundred people, that would have been about a 2.6 vote margin. A truly significant margin, in my mind, is more like 60/40. Also, nearly half of the eligible voters didn't even bother to show up.
Basically, what this election showed us is that A. many people could not tell much difference between the two candidates, so they voted to keep the guy with a four-year track record as President instead of voting in the guy with no real record in national politics, and B. among everyone who saw a difference between the two candidates, there were about as many people who liked/hated each candidate.
In other words, both candidates were remarkably bad at appealing to the masses.
They got a letter from me pointing out a few gaping flaws. The biggest ones were:
I concluded by saying that the purpose of reform should be ensuring that works aren't protected once they're no longer making money. In my mind, there's no benefit to mandating that every possible work is out of copyright in a given maximum number of years. If Disney or Lucas wants to pay a fortune to keep exclusive rights to Steamboat Willie, that's not really a problem so long as 99% of works go into the public domain in a timely manner.
In other words, it is a good start, but there are enough rough edges that the paper needs more review before it serves as any sort of starting point for legislation.
No, you're misunderstanding me. I don't have the right to be heard in that you have the right to not listen. I do have the right to be hearable, however. That's a crucial distinction.
For example, it's okay for a politician to buy up a bunch of first-adjacent or last-adjacent ad slots to air their attack ads. It's not okay for them to buy up every adjacent ad slot on every station and fill them all with pictures of poodles so that his or her opponent cannot get any air time at all.
To a point. Beyond a certain point, however, an organization of people can amass the ability to speak at sufficient volume (both in loudness and in quantity) to effectively drown out dissenting voices. Thus, in order to guarantee free speech for the individual, to some degree, the speech of large groups must be kept in check.
This applies whether the large group is the government, a PAC, a corporation, a union, or pretty much any other group of people who have united for a single purpose. For example, the U.S. government is structured very deliberately to minimize tyranny of the majority.
Similarly, commercial speech (e.g. advertising) has limits that require a degree of truthfulness. Commercial speech is not really different from any other speech, apart from the fact that the people doing the speaking have the presumption of authority and the volume to back up that presumption. We just have more laws to cover that corner case because it is fairly common and is often egregious. :-)
No, that's basically how FaceTime works, too. You call someone on the phone, and if the other end is an iOS device, there's a button you can push to switch the call over to FaceTime. It might require an iCloud account (I'm honestly not sure), but they're free, so that's not particularly important....
Have you looked into the wireless broadband carriers? I mean the ones with an antenna on the roof, not cellular. They cover most of the greater Silicon Valley area unless you're up in the mountains or something.
You do realize that you can choose the packaging, right? You can't usually choose whether to use PPPoE or PPPoA, for the most part. It's like your only shipper is a really greedy pack-and-ship firm that deliberately uses oversized boxes even for non-fragile items that are already in their own boxes.
I'm pretty sure you'll find that any OS less than a decade old supports path MTU discovery for IPv4. So that should only be an issue if you're running some really ancient OS.
The first step towards getting better is admitting you have a problem.
As I often say after tuning my horn, "Close enough for jazz."
Actually, he was only about half right. Used tastefully and in moderation, dissonance can create mood in ways that consonance cannot easily match. As with nearly all of the musical techniques that he argued were historically dissonant (with the exception of basic polyphony), however, used in excess, it sounds like crap.
IMO, the key to the tasteful use of dissonance is to make sure that the dissonance is not the focus. On the one extreme, you might have the subtle use of dissonant suspension and release in secondary parts of a complex orchestral work to set the mood. On the other extreme, you might have a highly dissonant piece of music used as the background sound behind a Civil War battle. In both cases, the listener is focused on something else, whether that something else is a traditional melodic line or a bunch of people shooting each other in a horrible, bloody battle.
Incidentally, most folks (statistically) don't like heavy metal, highly dissonant jazz, bebop, etc. even to this day. Those genres and subgenres all serve a useful purpose when it comes to expanding the musical universe, and over time, those experimental ideas will get incorporated into more mainstream music in much more subtle and toned-down ways, but that doesn't mean that most people will ever find the experimental music itself enjoyable to listen to.
And then on the flip side of dissonance, you have Charles Ives, whose music often has a tonal center. In fact, quite often, it has three or four.
If I were one of those folks, I would follow these steps to register a complaint with Apple. Just saying.
No, these days, I'm pretty sure the only way to survive getting a PhD in anything with your sanity intact involves a copious amount of the strong stuff....
At least it loads for you. In the current version of Safari (OS X), I just get:
TypeError: Attempting to change value of a readonly property. kanvas.js:49
Not at all. You idle in your driveway and at traffic lights for a couple of minutes per trip. An average processor while running a typical word processor or web browser is idle up to 98% of the time while waiting for the user to click or type something. Even if you drive on highways in bumper-to-bumper traffic every day, your engine's idle percentage isn't going to be that bad. And even when it is relatively non-idle, it usually isn't using more than 20-30% of a single CPU core unless you're playing a game, playing video, or doing some other similarly CPU-intensive task.
Measuring the power consumption with a benchmark is like measuring fuel economy at 120 MPH uphill all the way while towing a trailer. It isn't real-world use. When you measure a car's fuel economy, you measure it part of the time at low speed, part of the time at high speed, part of the time stoppping and starting, and part of the time idling, because that's how cars are actually driven. But even that isn't the way your computer is used.
Measuring CPU power usage correctly is more like measuring the fuel economy impact of a car's air conditioner. You don't measure it in 150 degree heat so that the air conditioner compressor runs 100% of the time, because that's not the way it is used in the real world. In the real world, the compressor runs for one minute and then disengages for ten.
And computers have always needed to be more efficient when idle. The only reason any of this stuff is happening is because Intel is starting to panic at the possibility of their share of the computing market drying up because of their idle wattage. Just look at Dell prototyping 64-bit ARM-based servers, and you'll understand why Intel is running scared.
If anything, idle power consumption matters even more for beefy hardware than for things like cell phones. You can always put a bigger battery in a cell phone and nobody is going to scream. When you're dealing with server farms that consume millions of dollars per year in electricity, significant reductions in idle power consumption can save a lot of money even if it means adding more hardware (so long as that extra hardware doesn't bring the idle power consumption back up to where it was before, that is). Granted, demand-based cluster scaling can help with power consumption, but those efficiency wins come at a cost in terms of the ability to rapidly respond to large changes in demand.
Haswell, assuming it lives up to the hype, should be a big step in the right direction, but since we're talking about hardware that (from what I've read, anyway) won't ship for several more months, and for which no public benchmark data exists AFAIK, it's really too soon to say for sure.
That's a false comparison, though. If users mostly ran benchmarks 24x7, that would be a good test of efficiency. The reality, however, is that CPUs mostly sit idle, so to compute average efficiency, you have to factor that in.
Granted, a faster CPU that can reach an idle state sooner can be more efficient than a slower CPU that runs at full bore for a longer period of time, but only if the idle wattage is fairly similar.