I sincerely don't think Germans think that way. These are the people who turned off all their nuclear plants for solar. I disagree with that action, but I have a lot of trouble believing that they think that the regulations of the US, of all places, are ridiculous. Germans are not known for their dislike of regulations.
Except that the regulations for small diesels are far stricter in the US than in Europe, so the Germans probably have good cause for thinking the US regulations are ridiculous.
And yet the undergrads pay a high tuition that funds said academics.
Hah! Of course it does, and researchers spend so much of their time chasing grants just for the fun of it.
Our research group got very little money from the university. Tenure salaries, maybe, and the occasional TA/RA position, but of course everyone was required to teach, self-funded or not.
You don't know what you're talking about if you think university is only about teaching undergrads. It isn't, and shouldn't be.
I'm arguing against OP's characterization of Linux users as tightwads, and never suggested that Linux had a large market share (will freely admit that's not the case). Don't know why you're suddenly going off in a different direction.
Go check the humble bundle website any time there's a deal on. They post statistics versus OS, and most of the time Linux comes out on top in terms of how much they pay. I'll admit I haven't checked the last few bundles, though.
Most Linux users probably do use free software, but that doesn't make them cheapskates, and HB's stats in fact show the opposite: when there's software that's *worth* paying for, Linux users voluntarily pay more.
Ok, so you're going to ignore the point of this discussion again - a discussion on the limitations of sharing spectrum, the main issue affecting wireless - and instead try to turn it into a rant on telecoms underinvesting? Guess we're done here then.
I see the point you're making now, but I would still differentiate the two oversubscription cases. In the case of ADSL, you can have oversubscription in two places: in the provider backbone, or in the bridge between the ADSL lines and the backbone. In the case of DOCSIS, there are three possible oversubscription points: the provider backbone, the bridge between the backbone and the DOCSIS network, and the shared coax cable itself.
To give you a more concrete illustration, if a provider were to keep adding new customers in a neighborhood, and kept upgrading their backbone and last-mile termination point in step with this, they'd still run into a brick wall when their coax saturates. There's no such wall with ADSL, precisely because coax is a shared last mile and twisted pair is not.
That solves the problem for some people, but not for those in my situation - I just moved across the country. People from other countries already face this problem: they have no US credit history, so for years they're screwed as far as credit is concerned.
I must admit that I don't know what to do about it, however. I can see the system is broken, but I don't know how to fix it. The European (well, Belgian as far as I know) solution is to not have credit history at all and instead to have far stricter bankruptcy laws, but those laws are a ball and chain around the country's entrepreneurs.
On the off chance you're not just trolling, let me ask you this: have you seriously never heard of the slowdown that happens pretty much every day in the evening with cable internet, and never with ADSL? If you have, do you seriously attribute that to the oversubscribing of the cable provider's backbone rather than DOCSIS?
I agree with your assessment, but shouldn't the FCC then be going after the radios, not the rest of the board? If the radio is licensed for channels 1-11, it shouldn't be capable of operating on channel 13 at all (also legal in the EU, btw). That would still allow us to flash the firmware without allowing the illegal operation you're talking about.
Essentially, installing DD-WRT should still not let you enable channel 13 in the US, but it should still be possible to install DD-WRT.
ADSL, the access technology connecting you to the DSLAM, is a point-to-point link. It is a dedicated channel for the last mile connecting a single subscriber to the provider backbone. DOCSIS, on the other hand, is by definition a shared medium from the moment it exits your modem onto the coax all the way to it's termination point where it connects to the backbone.
Yes, the backbone is shared, but that is entirely beside the point when we're discussing last-mile. Please do try to keep up with the discussion if you're going to contribute.
You're absolutely right that "multiple users" is the issue. You only have to look at DOCSIS, and what happens to that cable-based shared medium when everyone gets home from work and decides to start streaming. This is why ADSL can often offer better practical performance despite having a lower theoretical bandwidth. If you want consistent speed you need a dedicated channel.
Yes, you can achieve this with a microwave link. It's easier to provide the necessary isolation with a bounded cable medium rather than manipulating the antenna pattern, however, so the medium is not irrelevant.
Well, either the urban crowd *does* have a majority, regardless of the numbers GP pulled out of his ass, or the rural crowd is not as unified as GP thinks. Either way GP's an idiot for conflating land holding with voting rights, and I don't care enough to be precise.
Baby Jesus can go fsck himself, my math skills are fine.
Using 2012 as an example, Cook County contributed 1.94 million votes to a 5.1 million total. So 2.8% of the land area represented 40% of the results that decided 100% of the electoral votes of the state. I'm pretty sure the 97.2% of Illinois that works that land to feed the remaining 2.8% feels pretty crappy about that imbalance.
Yes, I'm sure those 97.2% of Illinois's acres of land feels really bad about supporting the other 2.8% of acres of land. On the other hand, the *minority* of rural voting humans don't have a leg to stand on when the *majority* of urban humans overrules them.
Do you want a state electoral college that choses the governor and legislators, with the electors chosen by each district, as they do at the federal level? Or maybe, say, one vote per person per acre owned?
Fair arguments, but those are arguments against self-driving cars in general as well. What do we do in general when the AI encounters a situation it can't handle? Although self-driving cars are apparently better than normal drivers in most conditions, it would surprise me if humans are not still much better at the edge cases. And edge cases that the car can't handle will always exist - what's the AI supposed to do when it faces a tornado, an earthquake, or a crazed gunman for instance?
Of course, there are also edge cases that a human driver can't handle, which is one of the reasons even good drivers still have accidents. It's impossible to train someone, be it a human or an AI, for all possible cases, so we just have to accept that limitation and train them for the cases they can reasonably be expected to encounter. With a self-driving car that would hopefully be a far smaller list, and would therefore require less certification.
Hopefully the off-switch approach would only be a stopgap measure until the AI is good enough to handle traffic cops and other edge cases, at which point anyone, capable or not, could hop in for a ride.
I'm not sure what you're getting at... Are you saying you're worried that car companies will say they're merely giving the cops the right to disable the self-driving mode of our car while *actually* giving them the power to transfer control of the self-driving AI to the cop?
Just in case my idea wasn't clear, I'm saying that the self-driving AI should *always* be under control of the person in the car, and that it should have nothing more for the cop than an off switch. In that scenario and assuming the car company doesn't screw it up, worst case is you have to drive manually, and your car can never be taken over by an external "driver".
Ok, so none of us likes the idea of the cops being able to take full control of the car. It leaves a security hole ripe for abuse and mischief, and the cop is not necessarily in the position to determine how best the car should move.
However, under the assumption that a self-driving car will have a manual mode, what if the cop could emit a signal that disabled the autopilot? That would put the driver in control again, who could then decide whether or not to follow the cop's instructions as well as determine how to do so in the safest manner. Make it a broadcast signal that blankets a certain area, so that the driver gets plenty of warning that they'll have to take over before arriving at the controlled intersection.
That should work long enough for us to figure out how to have an AI recognize a traffic cop's instructions.
This "businessman" has driven his various companies bankrupt on at least 3 occasions
Nothing wrong with declaring bankruptcy. If someone lends you money, that's done with the understanding that there's a chance you may not pay them back. That's one of the reasons you pay interest on lending, as otherwise there'd effectively be no cost (barring administrative) associated with lending and competitive pressure would then quickly push rates down to near zero.
No one succeeds all the time. This concept that you always have to pay back your debts no matter what is silly and counterproductive. One of the reasons the US is more successful in business than the EU is because the US is less harsh with debtors.
Multiple responses to the issue, but none of which, apart from a wasteful personal heater, are conducive to work. Tried mittens, also makes it slightly harder to type, which is an issue since that's what I do all day. I could "just" wear mittens about as easily as you could "just" wear more antiperspirant.
And why can't you have a personal AC? Unless you work in a closet with a 5 ft ceiling, the waste heat's just going to go up and away into the AC system. Or you could get a fan or dehumidifier, wear lighter fabrics, drink cold water...
As for 70-72, dunno, probably, possibly. my point is more that both sides have options (despite how GP and you portray things), and neither side likes them so tries to shift the responsibility onto the other. What we need is a new series of studies *and* more zonal control.
According to the carrier, if you want a secure phone, you'll just have to buy a new one from them.
Hang on a second... I understand what you're saying, and I'll definitely believe it applies to phones originally bought from a carrier. However, if I were to buy a Samsung directly from the manufacturer and then use it with a carrier, I'm not beholden to the carrier for updates, right? Since it's not a carrier-branded phone, I can just get updates over any valid internet connection, right?
Or does what you're saying even apply to non-branded unlocked phones? If it does, wow... I didn't realize the update regime was *that* screwed up.
The article title is somewhat misleading. If you watch the video on the Swincar, you'll see it's doing a lot more than just tilting. The front and rear "axel" can independently pivot around the roll axis of the car, and looks like each wheel can independently extend its reach. There's a photo of it driving straight forward, aligned straight in a gulch, and still with only its rear left and front right wheel in the gulch while the rear right and front left are on opposing hillsides of the gulch.
Finally, who actually thinks that people set temperature controls based on studies done in the 1960s?
Our building is. It was built in the 70's, doesn't have zonal control, so the entire building is set to somewhere around 68 F and the offices are hot or cold depending on where they are in the ducting. I don't have great circulation, so some days it's so chilly my hands start to stiffen and I can't type properly any more, which a sweater has limited effect on.
Yes, you may not be able to work when it's too hot, but I can't work when it's too cold, and it's not always as simple as just wearing more clothes. No one in my building wears a suit, and considering the weather's 90 F outside it would save them energy to turn the airco down, say to 75 F.
Depends on in which part of the country you are. I see a ~ $0.50/gal price difference in stations within 3 miles of home (Southern CA), and there's rarely a line at the cheaper places. It may only save me ~ $6 a tank, but that adds up over time, and the cheaper places are close enough that it's not really out of my way.
So explain how this corner case you are worried about trumps the far more common case of people who commit fraud asking for their records to be "forgotten" so they can continue to defraud people?
Because people who actually commit a crime have no right to be forgotten under this law. Your argument is not applicable.
...but you also get noise/junk audio when you layer multiple audio streams (drums, guitar, vocals, etc) and try to combine them into one waveform.
Can you please go into a bit more detail about this? Assuming you have enough headroom that you don't clip, and if you have a filter to remove any components above your final sample rate so they don't alias down, I'm not sure I understand where the junk would be coming from. Are you talking about applying non-linear effects to the sound?
I know that higher sampling rates improve things, in that it's at the very least easier to create good filters to crop audio to your final rate. This is important because any frequency components above 22 kHz are aliased down to audible frequencies when you downsample to 44 kHz, and it is far easier to build a 20 kHz digital low-pass filter running at a sample rate of 192 kHz than at, say, 48 kHz. Your quantization noise is also moved up to half the sample rate and can likewise be better filtered out than if it was at 22 kHz, which is what I meant by moving it around.
The noise floor drops directly from the higher bit depth
You mean distortion, right, and I suppose jitter as well? Thermal noise, which is always present, will go up.
Slashdot Mirror
Except that the regulations for small diesels are far stricter in the US than in Europe, so the Germans probably have good cause for thinking the US regulations are ridiculous.
Hah! Of course it does, and researchers spend so much of their time chasing grants just for the fun of it.
Our research group got very little money from the university. Tenure salaries, maybe, and the occasional TA/RA position, but of course everyone was required to teach, self-funded or not.
You don't know what you're talking about if you think university is only about teaching undergrads. It isn't, and shouldn't be.
I'm arguing against OP's characterization of Linux users as tightwads, and never suggested that Linux had a large market share (will freely admit that's not the case). Don't know why you're suddenly going off in a different direction.
Go check the humble bundle website any time there's a deal on. They post statistics versus OS, and most of the time Linux comes out on top in terms of how much they pay. I'll admit I haven't checked the last few bundles, though.
Most Linux users probably do use free software, but that doesn't make them cheapskates, and HB's stats in fact show the opposite: when there's software that's *worth* paying for, Linux users voluntarily pay more.
Ok, so you're going to ignore the point of this discussion again - a discussion on the limitations of sharing spectrum, the main issue affecting wireless - and instead try to turn it into a rant on telecoms underinvesting? Guess we're done here then.
I see the point you're making now, but I would still differentiate the two oversubscription cases. In the case of ADSL, you can have oversubscription in two places: in the provider backbone, or in the bridge between the ADSL lines and the backbone. In the case of DOCSIS, there are three possible oversubscription points: the provider backbone, the bridge between the backbone and the DOCSIS network, and the shared coax cable itself.
To give you a more concrete illustration, if a provider were to keep adding new customers in a neighborhood, and kept upgrading their backbone and last-mile termination point in step with this, they'd still run into a brick wall when their coax saturates. There's no such wall with ADSL, precisely because coax is a shared last mile and twisted pair is not.
That solves the problem for some people, but not for those in my situation - I just moved across the country. People from other countries already face this problem: they have no US credit history, so for years they're screwed as far as credit is concerned.
I must admit that I don't know what to do about it, however. I can see the system is broken, but I don't know how to fix it. The European (well, Belgian as far as I know) solution is to not have credit history at all and instead to have far stricter bankruptcy laws, but those laws are a ball and chain around the country's entrepreneurs.
On the off chance you're not just trolling, let me ask you this: have you seriously never heard of the slowdown that happens pretty much every day in the evening with cable internet, and never with ADSL? If you have, do you seriously attribute that to the oversubscribing of the cable provider's backbone rather than DOCSIS?
I agree with your assessment, but shouldn't the FCC then be going after the radios, not the rest of the board? If the radio is licensed for channels 1-11, it shouldn't be capable of operating on channel 13 at all (also legal in the EU, btw). That would still allow us to flash the firmware without allowing the illegal operation you're talking about.
Essentially, installing DD-WRT should still not let you enable channel 13 in the US, but it should still be possible to install DD-WRT.
ADSL, the access technology connecting you to the DSLAM, is a point-to-point link. It is a dedicated channel for the last mile connecting a single subscriber to the provider backbone. DOCSIS, on the other hand, is by definition a shared medium from the moment it exits your modem onto the coax all the way to it's termination point where it connects to the backbone.
Yes, the backbone is shared, but that is entirely beside the point when we're discussing last-mile. Please do try to keep up with the discussion if you're going to contribute.
You're absolutely right that "multiple users" is the issue. You only have to look at DOCSIS, and what happens to that cable-based shared medium when everyone gets home from work and decides to start streaming. This is why ADSL can often offer better practical performance despite having a lower theoretical bandwidth. If you want consistent speed you need a dedicated channel.
Yes, you can achieve this with a microwave link. It's easier to provide the necessary isolation with a bounded cable medium rather than manipulating the antenna pattern, however, so the medium is not irrelevant.
Well, either the urban crowd *does* have a majority, regardless of the numbers GP pulled out of his ass, or the rural crowd is not as unified as GP thinks. Either way GP's an idiot for conflating land holding with voting rights, and I don't care enough to be precise.
Baby Jesus can go fsck himself, my math skills are fine.
As opposed to the tyranny of the minority landholding aristocracy raiding the federal coffers?
Well, I'm glad you'll never get your way. Props for being honest, however.
Yes, I'm sure those 97.2% of Illinois's acres of land feels really bad about supporting the other 2.8% of acres of land. On the other hand, the *minority* of rural voting humans don't have a leg to stand on when the *majority* of urban humans overrules them.
Do you want a state electoral college that choses the governor and legislators, with the electors chosen by each district, as they do at the federal level? Or maybe, say, one vote per person per acre owned?
Fair arguments, but those are arguments against self-driving cars in general as well. What do we do in general when the AI encounters a situation it can't handle? Although self-driving cars are apparently better than normal drivers in most conditions, it would surprise me if humans are not still much better at the edge cases. And edge cases that the car can't handle will always exist - what's the AI supposed to do when it faces a tornado, an earthquake, or a crazed gunman for instance?
Of course, there are also edge cases that a human driver can't handle, which is one of the reasons even good drivers still have accidents. It's impossible to train someone, be it a human or an AI, for all possible cases, so we just have to accept that limitation and train them for the cases they can reasonably be expected to encounter. With a self-driving car that would hopefully be a far smaller list, and would therefore require less certification.
Hopefully the off-switch approach would only be a stopgap measure until the AI is good enough to handle traffic cops and other edge cases, at which point anyone, capable or not, could hop in for a ride.
I'm not sure what you're getting at... Are you saying you're worried that car companies will say they're merely giving the cops the right to disable the self-driving mode of our car while *actually* giving them the power to transfer control of the self-driving AI to the cop?
Just in case my idea wasn't clear, I'm saying that the self-driving AI should *always* be under control of the person in the car, and that it should have nothing more for the cop than an off switch. In that scenario and assuming the car company doesn't screw it up, worst case is you have to drive manually, and your car can never be taken over by an external "driver".
Ok, so none of us likes the idea of the cops being able to take full control of the car. It leaves a security hole ripe for abuse and mischief, and the cop is not necessarily in the position to determine how best the car should move.
However, under the assumption that a self-driving car will have a manual mode, what if the cop could emit a signal that disabled the autopilot? That would put the driver in control again, who could then decide whether or not to follow the cop's instructions as well as determine how to do so in the safest manner. Make it a broadcast signal that blankets a certain area, so that the driver gets plenty of warning that they'll have to take over before arriving at the controlled intersection.
That should work long enough for us to figure out how to have an AI recognize a traffic cop's instructions.
Nothing wrong with declaring bankruptcy. If someone lends you money, that's done with the understanding that there's a chance you may not pay them back. That's one of the reasons you pay interest on lending, as otherwise there'd effectively be no cost (barring administrative) associated with lending and competitive pressure would then quickly push rates down to near zero.
No one succeeds all the time. This concept that you always have to pay back your debts no matter what is silly and counterproductive. One of the reasons the US is more successful in business than the EU is because the US is less harsh with debtors.
Multiple responses to the issue, but none of which, apart from a wasteful personal heater, are conducive to work. Tried mittens, also makes it slightly harder to type, which is an issue since that's what I do all day. I could "just" wear mittens about as easily as you could "just" wear more antiperspirant.
And why can't you have a personal AC? Unless you work in a closet with a 5 ft ceiling, the waste heat's just going to go up and away into the AC system. Or you could get a fan or dehumidifier, wear lighter fabrics, drink cold water...
As for 70-72, dunno, probably, possibly. my point is more that both sides have options (despite how GP and you portray things), and neither side likes them so tries to shift the responsibility onto the other. What we need is a new series of studies *and* more zonal control.
Hang on a second... I understand what you're saying, and I'll definitely believe it applies to phones originally bought from a carrier. However, if I were to buy a Samsung directly from the manufacturer and then use it with a carrier, I'm not beholden to the carrier for updates, right? Since it's not a carrier-branded phone, I can just get updates over any valid internet connection, right?
Or does what you're saying even apply to non-branded unlocked phones? If it does, wow... I didn't realize the update regime was *that* screwed up.
The article title is somewhat misleading. If you watch the video on the Swincar, you'll see it's doing a lot more than just tilting. The front and rear "axel" can independently pivot around the roll axis of the car, and looks like each wheel can independently extend its reach. There's a photo of it driving straight forward, aligned straight in a gulch, and still with only its rear left and front right wheel in the gulch while the rear right and front left are on opposing hillsides of the gulch.
Our building is. It was built in the 70's, doesn't have zonal control, so the entire building is set to somewhere around 68 F and the offices are hot or cold depending on where they are in the ducting. I don't have great circulation, so some days it's so chilly my hands start to stiffen and I can't type properly any more, which a sweater has limited effect on.
Yes, you may not be able to work when it's too hot, but I can't work when it's too cold, and it's not always as simple as just wearing more clothes. No one in my building wears a suit, and considering the weather's 90 F outside it would save them energy to turn the airco down, say to 75 F.
Depends on in which part of the country you are. I see a ~ $0.50/gal price difference in stations within 3 miles of home (Southern CA), and there's rarely a line at the cheaper places. It may only save me ~ $6 a tank, but that adds up over time, and the cheaper places are close enough that it's not really out of my way.
Because people who actually commit a crime have no right to be forgotten under this law. Your argument is not applicable.
Can you please go into a bit more detail about this? Assuming you have enough headroom that you don't clip, and if you have a filter to remove any components above your final sample rate so they don't alias down, I'm not sure I understand where the junk would be coming from. Are you talking about applying non-linear effects to the sound?
I know that higher sampling rates improve things, in that it's at the very least easier to create good filters to crop audio to your final rate. This is important because any frequency components above 22 kHz are aliased down to audible frequencies when you downsample to 44 kHz, and it is far easier to build a 20 kHz digital low-pass filter running at a sample rate of 192 kHz than at, say, 48 kHz. Your quantization noise is also moved up to half the sample rate and can likewise be better filtered out than if it was at 22 kHz, which is what I meant by moving it around.
You mean distortion, right, and I suppose jitter as well? Thermal noise, which is always present, will go up.