Would have been interesting to look at whether improved customer satisfaction was correlated with increased local competition. I strongly suspect it is, not just because Comcast works harder to try to retain customers, but largely because the unhappiest customers leave as soon as they have an alternative. Even if actual customer service doesn't get any better, the people who remain are more satisfied on average.
Kind of like the famous demotivator says- "sometimes the best solution to workplace morale problems is just to fire all of the unhappy people."
When Google Fiber came to my area, Comcast sent door-to-door salespeople to try to get signups before people were committed to Google; Comcast was offering something like half its usual price. I knew it wasn't the salesperson's fault, but I couldn't help laughing in his face.
Most European countries rely heavily on sales taxes/ VAT. Such a tax does very little to distort people's incentives or discourage productive behavior. The result is that they have significant revenue (and, unless they're profligate, less deficit spending) with less deadweight loss to the economy.
In the US, we rely primarily on taxing productive behavior (payroll, savings, income, corporate taxes).
We also fill the tax code with enough loopholes and targeted cuts that it resembles a sieve. The targeted cuts are effectively government spending/subsidies; they may seem well intentioned in isolation but on the whole they're doing more harm than good. (Bush Sr.'s advisers had the motto "broaden the base, lower the rate," which is part of why some of the Clinton years ran a surplus. I imagine the loopholes were back in force by midway through Bush Jr's presidency.)
Even if we have lower taxes on the whole, in many cases our taxes are doing more harm to businesses and workers. We can change this.
Consumption taxes have seemed to be a third rail- an untouchable topic - in US politics, largely because by themselves they are regressive. But there are plenty of ways to implement an overall progressive tax system using them, like the "Bradford X tax."
Again, it's not that there aren't solutions - solutions which reasonable people on both sides of the aisle should find acceptable. It's that we can't scrounge up the political will and get elected representatives to act reasonably.
It's obvious why airlines overbook- it's a worthwhile gamble, given how frequently people can't make their flight.
But it's also obvious that if no one is taking them up on their compensation offers when the flight is overbooked, they aren't paying the social costs of their gamble, and so they're getting away with defrauding people.
The solution is obvious. Especially if people have already rightfully boarded the plane, they should only be removed voluntarily. Everyone on board turned down $800 compensation for missing the flight, but I'm sure somebody would have accepted $2,000 or less. If once in ten thousand flights nobody accepts an offer less than $20,000, the airline will just have to take that risk into account when they decide how much to overbook.
It doesn't take a great economist to come to this obvious conclusion; it was my immediate reaction and many others'. But I'll mention that a great economist has posted the same thing.
In this case, where it wasn't really overbooked but the airline needed to transport employees, already at $800 it's odd that they didn't just find another way to get one of their employees there (even by taxi).
I didn't have any special reason to love CodePlex, but I'm still sorry to see it and so many other such services go.
As nice as Github's features are, electing a single organization (inevitably with its own political agenda) as the planet's one source for development repos is a tremendously bad idea. Way too much concentration of power for abuse and way too low an organizational bus factor in case something goes wrong.
I've been pleased with the changes made here at Slashdot and at SourceForge since Dice sold them to BizX. SourceForge has a long way to go in regaining trust and catching up on features, but it's headed in the right direction. The changes they're making will help stem the exodus and I for one certainly hope it becomes good enough to be a real competitor to GitHub for new projects.
GPUs are still seeing notable performance increases because the problems you solve with a GPU are embarrassingly parallel. CPU progress has largely stalled because it's hard to get additional per-thread performance without clocking higher; the low-lying instruction level parallelism fruit is all gone. And the physics of the situation doesn't allow continuing to scale clock speeds the way they scaled from 1994-2002.
There are design related gains we know could be achieved without any new materials. In particular, clockless processors could be a huge jump forward. But designing a clockless processor is extremely difficult. A lot of the methods, tools, and engineering that has been developed over the last 50+ years to allow a relatively small team of people to manage the complexity of billions of transistors simply don't apply any more when you're dealing with a clockless processor.
Seven months since setting out to refresh the Mozilla brand experience, weâ(TM)ve reached the summit. Thousands of emails, hundreds of meetings, dozens of concepts, and three rounds of research later, we have something to share.
Yeah, right. If you want to show that you're a web firm moving the Internet forward and not just a SJW echo chamber that takes millions in search engines' ad revenue and turns it into mindless groupthink "brand experience" marketing baloney, try the following two steps as a start:
Fire everybody who wasted time on those thousands of emails and hundreds of meetings
Bring back Brendan Eich
I'm glad Mozilla is employing Xiph people for next-gen codec work. I struggle to think of any other way any of what they've done in the last four years has really benefited anyone.
I started using Mozilla as my main browser way back with M7 in 1999. I tried to spread the good word during the dark days of IE6 complete dominance. I trusted the organization. That trust has been destroyed.
The jump in intelligibility and voice quality going from 4kHz narrowband to 6kHz mediumband is big- probably bigger than going from mediumband to 20kHz fullband. The distinguishing features of many consonants are between 3.5 and 6 kHz.
Finding some way to take advantage of information beyond narrowband - even if not trying to encode much of it - could be a distinct advantage for a low bitrate codec over existing competition.
I guess it's impressive to get anything other than straight noise out of less than 1kbps. But I've wondered why Rowe hasn't focused more on quality at more moderate (e.g. 2-3kbps) bitrates rather than continuing to seek ways to trade away some quality for an ever lower bitrate. It's been a couple years since I tried it out and came to that conclusion; this looks like that trend has continued.
I couldn't get my encoded samples to sound nearly as good as the samples posted on the codec2 site. And it seemed like the second-lowest bitrate at the time (1400?) sounded essentially just as good as the highest (3200), which meant it wasn't making effective use of the additional bits. The quality jump between its highest mode and the lowest Opus mode (at 6kbps) was huge. (EVS would be a big jump over that.)
From what I understand, codec2's most prominent competition operates at 2.4kbps and up and sounds noticeably better at those rates than codec2 does.
(hit submit on this just before leaving this morning, didn't see till I got back there was an error.)
The Alhambra Decree etc weren't part of the Inquisition itself; that's changing the topic. There weren't 100,000 prisoners who died in the Inquisition; that's roughly the total number of cases they heard. Estimates vary but all the ones actually based on the historical data say less than 10,000.
Anyhow, I'm certainly not saying that the existence of Mao means the treatment of Jews in Spain over the centuries wasn't a problem.
I am saying your perspective is skewed. The data really don't show that the share of violent behavior that's associated with religion is large compared to the share of total social behavior associated with religion.
The result of this study says that people reporting a positive religious experience really have areas in their brain active that are active in other experiences people report as positive.
There's no more justification in history or in these studies for claims that religion is a disease that causes violence than there is for saying the same of commerce or love or any of a number of other basic human behaviors.
That guy may know Prolog but he doesn't know history. The vast majority of his citations are 19th-century Protestant anti-Catholic tracts, with one of his few 20th-century sources being a conspiracy-theorist type Baptist missionary writing in 1960.
Ultimately, the pairing of classical reward responses when hearing music with learning a smattering of music theory may indicate a brain mechanism for greater music appreciation. So what?
That's not "bypassing rational centers of the brain and creating a loop." It's simply "these people had a positive experience and there were ideas that were associated with that positive experience." If anything, the fact that brain regions which are active in moral reasoning were especially active in these people suggests the opposite of "bypassing rational centers."
You've conveniently ignored the actual data and results of their study entirely and instead taken a couple of speculative comments ("here's an idea, please fund us") out of context and twisted them.
The old baloney about religion being a primary cause of violence is a ridiculous urban legend. Ultimately you can trace the exaggerations back to centuries-old partisan tracts. Actual historians (e.g. Encyclopedia of Wars) find religiously motivated wars to be roughly 2% of the total death count.
If what you get out of the Shoah is that Hitler was right on both counts - Judaism is a disease, as is Christianity - there's something fundamentally wrong, not just with your understanding of history but with you.
The Inquisition killed about 3,000 people over the course of 350 years. (Secular courts, of course, killed people at a much faster rate.) For some perspective, the Great Leap Forward killed 30,000,000 people in 3 years.
That conclusion can no more come out of this research than could the idea that listening to music is an illness.
The research simply said that people reporting a positive experience showed activity in the reward centers of their brains. Big surprise! Hey, going outside in the sunshine activates the reward center of my brain, maybe that's an illness too.
The slashdot headline is there because people who are irrational and partisan want to ignore what the research actually said and use lies about it to bludgeon others. Your silly attempt to join the dogpile amounts to using your fame to act as a bully. It's shameful.
That's completely absurd. Abraham Lincoln's belief in God was very evident not only in what he said to the masses but in his personal behavior, especially during his presidency.
For at least the majority of his adult life he would have been more of a Deist than a Protestant. He firmly held on to the idea of a moral God who shaped and gave order to events. His views on Providence and predestination reflect his Calvinist upbringing. It's hard to say exactly what else he believed and when; he was a very private person, and from what hints we do have, his positions on doctrine seem to have been fluid.
An old statement about ICANN says that the ideas people were coming up with for international governance would lead to more censorship. I don't see a newer statement directly on the topic.
No wonder he's a target. Everybody wants to be him.
My personal favorite Bruce Schneier Fact: "Most people use passwords. Some people use passphrases. Bruce Schneier uses an epic passpoem, detailing the life and works of seven mythical Norse heroes."
The archaic ARMv6 architecture CPU in the original Pi is radically different from the ARMv7+NEON of the Pi2 or the ARMv8 of the Pi3. When the Pi2 was released you said the performance advantage of ARMv7 builds optimized for the Pi2 wasn't big enough to justify the complication of having a separate OS image. But after the introduction of the Pi3, as people migrate to newer Pis and the rest of the open source ARM world takes v7 and NEON for granted, don't the scales start to tip in favor of builds for modern processors?
Mathematica devs in particular have said that having to target such disparate architectures in a single binary prevents them from using a high-performance BLAS, which slows many kinds of algorithms down dramatically. And many multimedia codecs have had extensive NEON optimizations but these don't always get enabled at runtime on Pi2/3.
Why the hate for Hello? I get the privacy/security/why are we integrating a 3rd party extension/etc concerns with Pocket. But Hello seems like a more natural fit in the browser itself, especially with the WebRTC push.
Those points in the phase space where there exist objects like brains that have records of other parts of the universe, necessarily have records of the universe as it was (is) in less-entropic states, because the processes involved in forming those records happen along paths from less entropic to more entropic states.
No. If you treat all parts of phase space equally, the most likely paths to "brains with memory records" are paths from higher entropy states. Even though memory-filled brains or any other kind of "record" spontaneously appearing from disorder is very unlikely, it is more likely than their emerging from an even lower entropy and thus even less likely state.
Without other reasons to believe an arrow of time exists - a privileged region of phase space or asymmetric physical laws- the likeliest explanation of whatever "record" that appears to indicate a lower-entropy past is that the records are random fluctuations from higher-entropy states. And there are no reasons why "records" should not equally well actually record "previous" higher-entropy states. The region of phase space where I sit here with a neuron arrangement "remembering" my cold chocolate this morning spontaneously getting hot is just as large as the region where I sit here with neurons that "remember" my hot chocolate getting cold.
But without any other asymmetries in physical laws, we have every reason to think entropy should increase going backwards in time just as well as forwards.
Decreases in entropy aren't impossible, they're just vanishingly unlikely. If you have a state with less than maximal entropy, of course you expect its entropy to increase. But because the physical laws are symmetric, if you're treating all volumes of state space equally, you also expect the same thing if you replace t with -t. The most likely way for a low entropy state to arise is as an unlikely random fluctuation from a higher entropy state. A past with lower entropy than the present is just as unlikely as a future with lower entropy than the present.
So if you define past and future via entropy, any state with less than maximal entropy has no past and both -t and t count as future.
This doesn't match our way of understanding the macroscopic world. We're much more inclined to believe we had an astoundingly unlikely astoundingly low-entropy initial condition. But from a stat mech point of view, without more of an idea why we should treat volumes of phase space this unequally, this amounts to a deus ex machina. So the arrow of time is still a puzzle and this is part of why people are still looking for hints of time asymmetry.
I first saw this news on the BBC yesterday. The account is really funny, and the tweet they quote at the start of the article is nicely representative; I'll reproduce it here for those who haven't RTFA yet:
Vladimir Putin (@DarthPutinKGB) May 27, 2016
Arriving at Athens today:
Customs: Name?
Me: Vladimir Vladimirovich Putin
Customs: Occupation?
Me: No, this time i'm just here for 2 days
Slashdot Mirror
"I did it because of the kids! They called me Mr. Glass!"
Would have been interesting to look at whether improved customer satisfaction was correlated with increased local competition. I strongly suspect it is, not just because Comcast works harder to try to retain customers, but largely because the unhappiest customers leave as soon as they have an alternative. Even if actual customer service doesn't get any better, the people who remain are more satisfied on average.
Kind of like the famous demotivator says- "sometimes the best solution to workplace morale problems is just to fire all of the unhappy people."
When Google Fiber came to my area, Comcast sent door-to-door salespeople to try to get signups before people were committed to Google; Comcast was offering something like half its usual price. I knew it wasn't the salesperson's fault, but I couldn't help laughing in his face.
Most European countries rely heavily on sales taxes/ VAT. Such a tax does very little to distort people's incentives or discourage productive behavior. The result is that they have significant revenue (and, unless they're profligate, less deficit spending) with less deadweight loss to the economy.
In the US, we rely primarily on taxing productive behavior (payroll, savings, income, corporate taxes).
We also fill the tax code with enough loopholes and targeted cuts that it resembles a sieve. The targeted cuts are effectively government spending/subsidies; they may seem well intentioned in isolation but on the whole they're doing more harm than good. (Bush Sr.'s advisers had the motto "broaden the base, lower the rate," which is part of why some of the Clinton years ran a surplus. I imagine the loopholes were back in force by midway through Bush Jr's presidency.)
Even if we have lower taxes on the whole, in many cases our taxes are doing more harm to businesses and workers. We can change this.
Consumption taxes have seemed to be a third rail- an untouchable topic - in US politics, largely because by themselves they are regressive. But there are plenty of ways to implement an overall progressive tax system using them, like the "Bradford X tax."
We should also shift some of the burden of "productivity taxes" to Pigouvian taxes, which tax things that cause costs to society. A good example was the revenue-neutral carbon tax proposed by Republicans in Washington state (and shot down by Democrats because it didn't give them more money to advance their social agenda).
Again, it's not that there aren't solutions - solutions which reasonable people on both sides of the aisle should find acceptable. It's that we can't scrounge up the political will and get elected representatives to act reasonably.
It's obvious why airlines overbook- it's a worthwhile gamble, given how frequently people can't make their flight.
But it's also obvious that if no one is taking them up on their compensation offers when the flight is overbooked, they aren't paying the social costs of their gamble, and so they're getting away with defrauding people.
The solution is obvious. Especially if people have already rightfully boarded the plane, they should only be removed voluntarily. Everyone on board turned down $800 compensation for missing the flight, but I'm sure somebody would have accepted $2,000 or less. If once in ten thousand flights nobody accepts an offer less than $20,000, the airline will just have to take that risk into account when they decide how much to overbook.
It doesn't take a great economist to come to this obvious conclusion; it was my immediate reaction and many others'. But I'll mention that a great economist has posted the same thing.
In this case, where it wasn't really overbooked but the airline needed to transport employees, already at $800 it's odd that they didn't just find another way to get one of their employees there (even by taxi).
I didn't have any special reason to love CodePlex, but I'm still sorry to see it and so many other such services go.
As nice as Github's features are, electing a single organization (inevitably with its own political agenda) as the planet's one source for development repos is a tremendously bad idea. Way too much concentration of power for abuse and way too low an organizational bus factor in case something goes wrong.
I've been pleased with the changes made here at Slashdot and at SourceForge since Dice sold them to BizX. SourceForge has a long way to go in regaining trust and catching up on features, but it's headed in the right direction. The changes they're making will help stem the exodus and I for one certainly hope it becomes good enough to be a real competitor to GitHub for new projects.
I, too, have partnered with Google, am your guest, and will share my tale.
Once upon a time, in the far-off land of Nigeria, there lived a prince.
This prince had inherited a fortune of FIFTEEN MILLION UNITED STATES DOLLARS...
GPUs are still seeing notable performance increases because the problems you solve with a GPU are embarrassingly parallel. CPU progress has largely stalled because it's hard to get additional per-thread performance without clocking higher; the low-lying instruction level parallelism fruit is all gone. And the physics of the situation doesn't allow continuing to scale clock speeds the way they scaled from 1994-2002.
There are design related gains we know could be achieved without any new materials. In particular, clockless processors could be a huge jump forward. But designing a clockless processor is extremely difficult. A lot of the methods, tools, and engineering that has been developed over the last 50+ years to allow a relatively small team of people to manage the complexity of billions of transistors simply don't apply any more when you're dealing with a clockless processor.
Yeah, right. If you want to show that you're a web firm moving the Internet forward and not just a SJW echo chamber that takes millions in search engines' ad revenue and turns it into mindless groupthink "brand experience" marketing baloney, try the following two steps as a start:
I'm glad Mozilla is employing Xiph people for next-gen codec work. I struggle to think of any other way any of what they've done in the last four years has really benefited anyone.
I started using Mozilla as my main browser way back with M7 in 1999. I tried to spread the good word during the dark days of IE6 complete dominance. I trusted the organization. That trust has been destroyed.
The jump in intelligibility and voice quality going from 4kHz narrowband to 6kHz mediumband is big- probably bigger than going from mediumband to 20kHz fullband. The distinguishing features of many consonants are between 3.5 and 6 kHz.
Finding some way to take advantage of information beyond narrowband - even if not trying to encode much of it - could be a distinct advantage for a low bitrate codec over existing competition.
The github mirror has a nicer interface.
I guess it's impressive to get anything other than straight noise out of less than 1kbps. But I've wondered why Rowe hasn't focused more on quality at more moderate (e.g. 2-3kbps) bitrates rather than continuing to seek ways to trade away some quality for an ever lower bitrate. It's been a couple years since I tried it out and came to that conclusion; this looks like that trend has continued.
I couldn't get my encoded samples to sound nearly as good as the samples posted on the codec2 site. And it seemed like the second-lowest bitrate at the time (1400?) sounded essentially just as good as the highest (3200), which meant it wasn't making effective use of the additional bits. The quality jump between its highest mode and the lowest Opus mode (at 6kbps) was huge . (EVS would be a big jump over that.)
From what I understand, codec2's most prominent competition operates at 2.4kbps and up and sounds noticeably better at those rates than codec2 does.
(hit submit on this just before leaving this morning, didn't see till I got back there was an error.)
The Alhambra Decree etc weren't part of the Inquisition itself; that's changing the topic. There weren't 100,000 prisoners who died in the Inquisition; that's roughly the total number of cases they heard. Estimates vary but all the ones actually based on the historical data say less than 10,000.
Anyhow, I'm certainly not saying that the existence of Mao means the treatment of Jews in Spain over the centuries wasn't a problem.
I am saying your perspective is skewed. The data really don't show that the share of violent behavior that's associated with religion is large compared to the share of total social behavior associated with religion.
The result of this study says that people reporting a positive religious experience really have areas in their brain active that are active in other experiences people report as positive.
There's no more justification in history or in these studies for claims that religion is a disease that causes violence than there is for saying the same of commerce or love or any of a number of other basic human behaviors.
That guy may know Prolog but he doesn't know history. The vast majority of his citations are 19th-century Protestant anti-Catholic tracts, with one of his few 20th-century sources being a conspiracy-theorist type Baptist missionary writing in 1960.
Ultimately, the pairing of classical reward responses when hearing music with learning a smattering of music theory may indicate a brain mechanism for greater music appreciation. So what?
That's not "bypassing rational centers of the brain and creating a loop." It's simply "these people had a positive experience and there were ideas that were associated with that positive experience." If anything, the fact that brain regions which are active in moral reasoning were especially active in these people suggests the opposite of "bypassing rational centers."
You've conveniently ignored the actual data and results of their study entirely and instead taken a couple of speculative comments ("here's an idea, please fund us") out of context and twisted them.
The old baloney about religion being a primary cause of violence is a ridiculous urban legend. Ultimately you can trace the exaggerations back to centuries-old partisan tracts. Actual historians (e.g. Encyclopedia of Wars) find religiously motivated wars to be roughly 2% of the total death count.
If what you get out of the Shoah is that Hitler was right on both counts - Judaism is a disease, as is Christianity - there's something fundamentally wrong, not just with your understanding of history but with you.
The Inquisition killed about 3,000 people over the course of 350 years. (Secular courts, of course, killed people at a much faster rate.) For some perspective, the Great Leap Forward killed 30,000,000 people in 3 years.
That conclusion can no more come out of this research than could the idea that listening to music is an illness.
The research simply said that people reporting a positive experience showed activity in the reward centers of their brains. Big surprise! Hey, going outside in the sunshine activates the reward center of my brain, maybe that's an illness too.
The slashdot headline is there because people who are irrational and partisan want to ignore what the research actually said and use lies about it to bludgeon others. Your silly attempt to join the dogpile amounts to using your fame to act as a bully. It's shameful.
That's completely absurd. Abraham Lincoln's belief in God was very evident not only in what he said to the masses but in his personal behavior, especially during his presidency.
For at least the majority of his adult life he would have been more of a Deist than a Protestant. He firmly held on to the idea of a moral God who shaped and gave order to events. His views on Providence and predestination reflect his Calvinist upbringing. It's hard to say exactly what else he believed and when; he was a very private person, and from what hints we do have, his positions on doctrine seem to have been fluid.
An old statement about ICANN says that the ideas people were coming up with for international governance would lead to more censorship. I don't see a newer statement directly on the topic.
It's common knowledge that if you knock out Chuck Norris with a roundhouse kick you become the new Chuck Norris.
Similarly, if you manage to steal Bruce Schneier's identity, you become the new Bruce Schneier.
No wonder he's a target. Everybody wants to be him.
My personal favorite Bruce Schneier Fact: "Most people use passwords. Some people use passphrases. Bruce Schneier uses an epic passpoem, detailing the life and works of seven mythical Norse heroes."
Thank you, come again.
The archaic ARMv6 architecture CPU in the original Pi is radically different from the ARMv7+NEON of the Pi2 or the ARMv8 of the Pi3. When the Pi2 was released you said the performance advantage of ARMv7 builds optimized for the Pi2 wasn't big enough to justify the complication of having a separate OS image. But after the introduction of the Pi3, as people migrate to newer Pis and the rest of the open source ARM world takes v7 and NEON for granted, don't the scales start to tip in favor of builds for modern processors?
Mathematica devs in particular have said that having to target such disparate architectures in a single binary prevents them from using a high-performance BLAS, which slows many kinds of algorithms down dramatically. And many multimedia codecs have had extensive NEON optimizations but these don't always get enabled at runtime on Pi2/3.
Why the hate for Hello? I get the privacy/security/why are we integrating a 3rd party extension/etc concerns with Pocket. But Hello seems like a more natural fit in the browser itself, especially with the WebRTC push.
Perhaps you have in mind the classic O. Henry story "The Ransom of Red Chief."
No. If you treat all parts of phase space equally, the most likely paths to "brains with memory records" are paths from higher entropy states. Even though memory-filled brains or any other kind of "record" spontaneously appearing from disorder is very unlikely, it is more likely than their emerging from an even lower entropy and thus even less likely state.
Without other reasons to believe an arrow of time exists - a privileged region of phase space or asymmetric physical laws- the likeliest explanation of whatever "record" that appears to indicate a lower-entropy past is that the records are random fluctuations from higher-entropy states. And there are no reasons why "records" should not equally well actually record "previous" higher-entropy states. The region of phase space where I sit here with a neuron arrangement "remembering" my cold chocolate this morning spontaneously getting hot is just as large as the region where I sit here with neurons that "remember" my hot chocolate getting cold.
But without any other asymmetries in physical laws, we have every reason to think entropy should increase going backwards in time just as well as forwards.
Decreases in entropy aren't impossible, they're just vanishingly unlikely. If you have a state with less than maximal entropy, of course you expect its entropy to increase. But because the physical laws are symmetric, if you're treating all volumes of state space equally, you also expect the same thing if you replace t with -t. The most likely way for a low entropy state to arise is as an unlikely random fluctuation from a higher entropy state. A past with lower entropy than the present is just as unlikely as a future with lower entropy than the present.
So if you define past and future via entropy, any state with less than maximal entropy has no past and both -t and t count as future.
This doesn't match our way of understanding the macroscopic world. We're much more inclined to believe we had an astoundingly unlikely astoundingly low-entropy initial condition. But from a stat mech point of view, without more of an idea why we should treat volumes of phase space this unequally, this amounts to a deus ex machina. So the arrow of time is still a puzzle and this is part of why people are still looking for hints of time asymmetry.
I first saw this news on the BBC yesterday. The account is really funny, and the tweet they quote at the start of the article is nicely representative; I'll reproduce it here for those who haven't RTFA yet: