Is there any evidence that offsets are actually effective in either preventing carbon from being emmitted and/or actively removing carbon from the atmosphere and putting it somewhere it won't be re-released for millenia?
but generally the prices are close once you start matching spec for spec.
The thing is that isn't how one normally buys computers, normally one starts with a set of requirements and then looks for a computer to meet those specs.
And when looked at in this way for many sets of requirements the cheapest mac that meets them is a LOT more expensive than the cheapest PC that meets them.
mmm, I find it sad, on the one hand I want to play the big hit games and I want to reward the developers for what they have created (I don't want to pirate stuff). OTOH I find the direction the gaming market is going with forced firmware updates on consoles and online activation (or worse) on the PC very unattractive.
If anything the XBOX seems to be the lesser of three evils at the moment, afaict they aren't requiring online activation (though they are taking steps towards it with in-box DLC) and afaict their firmware updates don't retroactively remove functionality.
I hope an actual customer friendly option comes out of this but I wouldn't hold my breath.
The SoC on this board has an internal ADC along with timers, PWM, and GPIOs. Not sure what you mean about latencies and jitter. Shouldn't a 32-bit CPU handle code and data accesses better than an 8-bit CPU? I'm not really a CPU architecture guy, so please correct me if I'm wrong.
In general
A simple CPU is a predictable CPU A fast CPU is a complex CPU
As you move up to more complex CPUs there gets to be more complexity between the core and the perhiperals. That complexity (caches, pipelining etc) causes the timing of your code to be harder to predict. Things get even harder to predict when you try to run an operating system or when you have internally generated non-maskable intterupts.
Still I don't think it's too big a deal for most applications.
I don't see any reason why you can't use an ARM SoC as the core of a hobbyist platform. It might not always be the best choice, but it does seem like a valid choice.
The armel designation is used to indicate little-endian ARM targets
When it comes to debian arm architectures things are something of a mess. While armel literally means "arm little endian" it is also the name of a debian architecture and in that context it says a lot more than just little endian.
There have been a number of different debian arm ports with different names
"arm" was little endian old abi with parameter passing done in a way that depends on an old FPU (which is kernel-emulated for systems that don't have it resulting in APPALLING floating point performance on most systems). It was an official port up until lenny and is now dead (other than security updates for lenny). "armeb" was big endian old abi intended for the NSLU2 and similar, was never official and IIRC died out sometime during the sarge era after people managed to get "arm" running on the NSLU2. There was talk of an eabi big endian port under the same name (since armeb old-abi was never an official port and is long dead the lack of compatibility wouldn't have been too big a deal) but nothing ever came of it. "armel" is little endian eabi with parameter passing done in a way that doesn't use any FPU (debian official binaries don't use any FPU at all but IIRC it IS possible to build binaries that use the armel ABI and yet make use of the FPU). It has been an official port since lenny. "armhf" is little endian eabi with parameter passing that depends on what is now the most common FPU. Afaict it's not currently an official port though there is some chance of it being in the future.
USB didn't kill anything. Firewire was never designed or intended to be the ubiquitous all-in-one connector.
Firewire was designed to be a high speed counterpart to USB, the idea was that your keyboard and mouse would still be USB but your high speed devices (high res scanners, external HDDs etc) would be firewire. Then the USB guys decided to release high-speed USB with a higher headline speed than firewire at a much lower cost. The result was that firewire was relegated to a handful of niches which actually needed it's better real world throughput and/or it's lower latency.
high speed USB "killed" firewire in the same sense that x86-64 "killed" itanium. It's not totally dead but it's relegated to a small fraction of it's original aspirations.
Laptop "docking stations" seem like the obvious use to me. Manufacturers have tried to make USB docking stations before with some success but USB2 isn't really fast enough to properly drive a display. USB3 is theoretically fast enough to push 1920x1080x60x24 but it's not really connected correctly internally (all the video data would have to be copied back over the systems general bus infrastructure to get it to the USB3 port) and i'm not sure it can acheive that data rate stably and reliably.
Thunderbolt is designed from the start to carry video and PCIe so afaict a very good docking station could be built simply by hooking a thunderbolt controller to a few PCIe based devices (e.g. a couple of USB3 controllers, a firewire controller and a gigabit ethernet controller)
The question is how much will such a docking station cost and will people consider that cost worthwhile to reduce the number of cables needed to hook up a laptop from 3 (power, USB and monitor) to two (power and thunderbolt)?
All Intel chipsets are scheduled to have Thunderbolt controllers in them, beginning with Ivy Bridge.
Do you have a source for that claim? anandtech say "Though some rumors reported Panther Point would include support for Thunderbolt, there is absolutely nothing in the current roadmap to suggest its presence in the 7-series chipsets"
BS eSata performs better but it has issues of it's own
1: the original connector didn't supply power, there is now a variant that combines eSata with power and USB signals on the same connector but i'm not sure how standardised it is. 2: on many machines the esata port is non-functional and/or non-hotpluggable because the BIOS was put in IDE mode to make installing windows easier. There is also usually only one eSata port. 3: only one layer of port multipliers is allowed and not all host controllers even support that. I've also never seen anyone selling a standalone eSata port multiplier only external to internal multipliers intended for integration into drive cabinets.
In short eSata is great if you just want to hook up ONE external drive to a few machines that you control, it's no so great for a drive you carry around between multiple machines that have other people administering them. You can get drives that do both USB and eSata but you still have to deal with multiple cables.
When the macbook (non-pro) went unibody they did drop firewire but the old non-unibody macbook was kept available until after the 13-inch macbook pro (which does have firewire) came out so there was always a 13 inch laptop in the range with firewire.
As for the other machines afaict:
The macbook air never had firewire. The macbook pros have always had firewire The mac pros have always had firewire The imacs have had firewire continuously since the imac G4
I'd hardly call that "abandoned" (BTW you do know that firewire 400 and firewire 800 are electrically compatible right? you just need the right cable)..
As in, if Bob spends some hours playing one particular game, and Dave spends those same hours sitting arround doing nothing, and then they both go get an all you can eat lunch, Bob will eat substantially more than Dave, all other things being equal.
There fixed it for you.
They then post a conclusion which extrapolates these very specific results to the general case.
Right so they take a very specific set of circumstances and show that in those circumstances among a small (and most likely far from random) sample set that those who played video games at more than those who sat arround doign nothing.
Then that gets generalised to "A single session of video game play in healthy male adolescents is associated with an increased food intake, regardless of appetite sensations. " which the media translates to "Video Game playing increases food intake in teens". Those are very bold general statements to make from a study that like many studies leaves more questions than answers. Notablly:
1: the duration of game playing and the time of food intake was fixed. While this may be representive of the situation in some families it's certainly not the situation everywhere. What difference does it make if the length of the activity is varied? What difference does it make if they have free access to food DURING the activity as well as after it. I know I tend to eat a LOT when i'm hanging arround doing nothing. 2: the test was with one specific game, there are many different intensities of game which one could reasonablly expect to have different impacts. 3: there was no comparision with substitute activities only with sitting arround doing nothing. The tests were also all proceeded by sitting arround doing nothing. 4: the measure was done with a rather early lunch, i'd think that apart from weekends most teenagers would be doing their game playing either between school and tea or after tea in the evening. 5: they were given an all you can eat lunch immediately after playing, does the impact happen only if play is immediately before free access to food or does it hang arround?
And that is the rub, the main customers left for high end* calculators are students. They buy them because they aren't allowed to take laptops, smartphones etc into exams but in at least some institutions they are allowed to take in particular models of high end calculator.
However high end calculators can also be used as cheating devices. Information that the students are supposed to remember can often be stored as can programs to do things the students are supposed to be able to do manually. Calculator vendors have responsed by introducing "exam modes" where the user is denied access to all memories and with some visual indication to allow invigilators to check the device is in exam mode. However this breaks down if cheaters can replace the firmware with one that pretends to be in exam mode but doesn't actually enforce the restrictions.
My department at uni doesn't allow high end calculators in exams at all . Personally I think that is the way to go as it's perfectly possible to test what needs to be tested without any questions that require a high end calculator to answer. The calculator vendors certainly wouldn't like that endgame though.
* by which I mean anything with more features than a basic "scientific" calculator.
Is that bandwidth is being shifted from one medium to another through the same output device.
Afaict the main difference is in where that bandwidth is from and to.
With traditional TV (whether delivered over cable, sattelite or terrestrial) the bandwidth is used in a broadcast manner. So on each network segment it's only used once per channel no matter how many users tune into that channel.
With ondemand provided by the last mile communications provider the communications provider has full control over where the bandwidth is to.
With internet TV they can ask the provider nicely to locate their servers more locally but without getting into nasty blocking practices there is little they can do to make them do so and managing n sets of third party servers in each location you want content to be streamed from is going to be a lot more overhead than maintaining one server that you control.
"Last mile" in telecoms reffers to the final connection from communications provider infrastructure to the use. In a big city it may be less than a mile in the countryside it may be significantly more.
Having said that while netflix traffic may not be running accross core internet backbones in signficant quantities I suspect it is going a lot further than the "last mile" connection in many if not most cases.
Some believe that there should be an online money transfer mechanism which offers similar levels of anonymity to cash and like cash can't easilly be blocked by governements. Many governements would rather such a system didn't exist and have legislated accordingly (for example with laws requiring money transfer buisnesses to collect details of their customers, freeze accounts etc).
Bitcoin was designed to be highly distributed and very difficult (though probablly not impossible) to subvert. That is it's attraction.
That and they are actually promoted, a great machine is of little use if noone knows it exists.
I have a HP mini 5101 with the high res screen option and I love the thing for giving me a usable screen resoloution (which IMO was the biggest issue with netbooks for general desktop tasks) in a small package but afaict HP never marketed it. So the only people who would have been likely to end up with one are those carefully researching the market for a machine with those charactersitcs.
MS has integrated instruction set emulation into windows before (they certainly did it on alpha and I think they did it on itanium too since the hardware level emulation was appalling and IIRC it was even removed in later versions of the itanium line), there is no reason they couldn't do it again. Intel are claiming they won't but then it's in intels interests to claim that given that they are trying to compete with ARM.
1: PC game vendors have been cracking down hard on the used game market ever since valve showed the industry they could get away with it. 2: Console game vendors have been taking more "baby steps". They still allow resale for the moment but they increasingly include content that can only be registered to one account.
They could do it steam style, ship a copy of the game encrypted on physical media and then force you to tie the key to your account before you can decrypt and install it. That way they could lock things to a purchaser while keeping bandwidth use low. Especially if they don't insist on updates (IIRC steam did insist on updating to the latest version as part of an install from physical media but I may be wrong there).
Afaict microsims are already as small as microSD, making them even smaller will make them very awkward to handle.
Not a big deal for those who live in one country and use one SIM all the time but a very big deal for those who travel and want to swap sims to get a deal that is decent locally.
I was under the impression that pretty much all VGA cables used a coax based design though on the cheaper cables it's probablly not a very high precision one. Certainly the ones i've cut open on old monitors did (I notice that LCDs tend to come with much thinner cables than the CRTs of old did, I guess they get away with a lower quality cable because of the short run length and the fact they are sampling the signal anyway rather than treating it as a continuous analog signal but I have never cut one of those cables open)
A perfect high frequency signal cable would have a core of perfect conductor surrounded by a perfectly dimensioned* insulator enclosed in a perfectly round and perfectly conductive screen. Of course that perfection is unatainable but the more you pay for your cable the closer they can get to those ideals, particularly the tolerances of the various dimensions and the quality of the screen. Raising the size of the cable (the relative sizes of different components of the cable are constrained by impedance requirements but the design as a whole can be scaled pretty freely) also tends to help both by reducing resistive losses and reducing the impact of sizing errors.
There are standards from the bodies that define HDMI. Since HDMI cables aren't a threat to safety or other critical things and afaict most cables do a pretty good job of complying I don't see any reason to make those standards mandatory.
The reason we have overpriced HDMI cables is twofold
1: Some people belive that more expensive always means better and don't get that with a digital cable things tend to either work perfectly or very obviously not work. 2: big electronics stores operate on the principle of being competitive on the big ticket items while overcharging on the extras. I don't particually like this practice but I don't see any other real way they could compete with online only vendors and I wouldn't want physical stores to dissapear.
Slashdot Mirror
Is there any evidence that offsets are actually effective in either preventing carbon from being emmitted and/or actively removing carbon from the atmosphere and putting it somewhere it won't be re-released for millenia?
but generally the prices are close once you start matching spec for spec.
The thing is that isn't how one normally buys computers, normally one starts with a set of requirements and then looks for a computer to meet those specs.
And when looked at in this way for many sets of requirements the cheapest mac that meets them is a LOT more expensive than the cheapest PC that meets them.
The same can be said for Microsoft and Apple.
mmm, I find it sad, on the one hand I want to play the big hit games and I want to reward the developers for what they have created (I don't want to pirate stuff). OTOH I find the direction the gaming market is going with forced firmware updates on consoles and online activation (or worse) on the PC very unattractive.
If anything the XBOX seems to be the lesser of three evils at the moment, afaict they aren't requiring online activation (though they are taking steps towards it with in-box DLC) and afaict their firmware updates don't retroactively remove functionality.
I hope an actual customer friendly option comes out of this but I wouldn't hold my breath.
The SoC on this board has an internal ADC along with timers, PWM, and GPIOs. Not sure what you mean about latencies and jitter. Shouldn't a 32-bit CPU handle code and data accesses better than an 8-bit CPU? I'm not really a CPU architecture guy, so please correct me if I'm wrong.
In general
A simple CPU is a predictable CPU
A fast CPU is a complex CPU
As you move up to more complex CPUs there gets to be more complexity between the core and the perhiperals. That complexity (caches, pipelining etc) causes the timing of your code to be harder to predict. Things get even harder to predict when you try to run an operating system or when you have internally generated non-maskable intterupts.
Still I don't think it's too big a deal for most applications.
I don't see any reason why you can't use an ARM SoC as the core of a hobbyist platform. It might not always be the best choice, but it does seem like a valid choice.
Agreed
The armel designation is used to indicate little-endian ARM targets
When it comes to debian arm architectures things are something of a mess. While armel literally means "arm little endian" it is also the name of a debian architecture and in that context it says a lot more than just little endian.
There have been a number of different debian arm ports with different names
"arm" was little endian old abi with parameter passing done in a way that depends on an old FPU (which is kernel-emulated for systems that don't have it resulting in APPALLING floating point performance on most systems). It was an official port up until lenny and is now dead (other than security updates for lenny).
"armeb" was big endian old abi intended for the NSLU2 and similar, was never official and IIRC died out sometime during the sarge era after people managed to get "arm" running on the NSLU2. There was talk of an eabi big endian port under the same name (since armeb old-abi was never an official port and is long dead the lack of compatibility wouldn't have been too big a deal) but nothing ever came of it.
"armel" is little endian eabi with parameter passing done in a way that doesn't use any FPU (debian official binaries don't use any FPU at all but IIRC it IS possible to build binaries that use the armel ABI and yet make use of the FPU). It has been an official port since lenny.
"armhf" is little endian eabi with parameter passing that depends on what is now the most common FPU. Afaict it's not currently an official port though there is some chance of it being in the future.
USB didn't kill anything. Firewire was never designed or intended to be the ubiquitous all-in-one connector.
Firewire was designed to be a high speed counterpart to USB, the idea was that your keyboard and mouse would still be USB but your high speed devices (high res scanners, external HDDs etc) would be firewire. Then the USB guys decided to release high-speed USB with a higher headline speed than firewire at a much lower cost. The result was that firewire was relegated to a handful of niches which actually needed it's better real world throughput and/or it's lower latency.
high speed USB "killed" firewire in the same sense that x86-64 "killed" itanium. It's not totally dead but it's relegated to a small fraction of it's original aspirations.
Laptop "docking stations" seem like the obvious use to me. Manufacturers have tried to make USB docking stations before with some success but USB2 isn't really fast enough to properly drive a display. USB3 is theoretically fast enough to push 1920x1080x60x24 but it's not really connected correctly internally (all the video data would have to be copied back over the systems general bus infrastructure to get it to the USB3 port) and i'm not sure it can acheive that data rate stably and reliably.
Thunderbolt is designed from the start to carry video and PCIe so afaict a very good docking station could be built simply by hooking a thunderbolt controller to a few PCIe based devices (e.g. a couple of USB3 controllers, a firewire controller and a gigabit ethernet controller)
The question is how much will such a docking station cost and will people consider that cost worthwhile to reduce the number of cables needed to hook up a laptop from 3 (power, USB and monitor) to two (power and thunderbolt)?
All Intel chipsets are scheduled to have Thunderbolt controllers in them, beginning with Ivy Bridge.
Do you have a source for that claim? anandtech say "Though some rumors reported Panther Point would include support for Thunderbolt, there is absolutely nothing in the current roadmap to suggest its presence in the 7-series chipsets"
USB 3 offers no advantages over eSATA.
BS eSata performs better but it has issues of it's own
1: the original connector didn't supply power, there is now a variant that combines eSata with power and USB signals on the same connector but i'm not sure how standardised it is.
2: on many machines the esata port is non-functional and/or non-hotpluggable because the BIOS was put in IDE mode to make installing windows easier. There is also usually only one eSata port.
3: only one layer of port multipliers is allowed and not all host controllers even support that. I've also never seen anyone selling a standalone eSata port multiplier only external to internal multipliers intended for integration into drive cabinets.
In short eSata is great if you just want to hook up ONE external drive to a few machines that you control, it's no so great for a drive you carry around between multiple machines that have other people administering them. You can get drives that do both USB and eSata but you still have to deal with multiple cables.
When the macbook (non-pro) went unibody they did drop firewire but the old non-unibody macbook was kept available until after the 13-inch macbook pro (which does have firewire) came out so there was always a 13 inch laptop in the range with firewire.
As for the other machines afaict:
The macbook air never had firewire.
The macbook pros have always had firewire
The mac pros have always had firewire
The imacs have had firewire continuously since the imac G4
I'd hardly call that "abandoned" (BTW you do know that firewire 400 and firewire 800 are electrically compatible right? you just need the right cable)..
Are there actually native USB3 hard drives out there or are they all just built on a USB3 to SATA bridge?
I'd expect
Southbridge->eSATA link->SSD
to have lower overheads than
Southbridge->PCIe (2.0 if you are lucky 1.0 on slightly older boards) x1 link->USB 3 controller-> Sata link->SSD
As in, if Bob spends some hours playing one particular game, and Dave spends those same hours sitting arround doing nothing, and then they both go get an all you can eat lunch, Bob will eat substantially more than Dave, all other things being equal.
There fixed it for you.
They then post a conclusion which extrapolates these very specific results to the general case.
Right so they take a very specific set of circumstances and show that in those circumstances among a small (and most likely far from random) sample set that those who played video games at more than those who sat arround doign nothing.
Then that gets generalised to "A single session of video game play in healthy male adolescents is associated with an increased food intake, regardless of appetite sensations. " which the media translates to "Video Game playing increases food intake in teens". Those are very bold general statements to make from a study that like many studies leaves more questions than answers. Notablly:
1: the duration of game playing and the time of food intake was fixed. While this may be representive of the situation in some families it's certainly not the situation everywhere. What difference does it make if the length of the activity is varied? What difference does it make if they have free access to food DURING the activity as well as after it. I know I tend to eat a LOT when i'm hanging arround doing nothing.
2: the test was with one specific game, there are many different intensities of game which one could reasonablly expect to have different impacts.
3: there was no comparision with substitute activities only with sitting arround doing nothing. The tests were also all proceeded by sitting arround doing nothing.
4: the measure was done with a rather early lunch, i'd think that apart from weekends most teenagers would be doing their game playing either between school and tea or after tea in the evening.
5: they were given an all you can eat lunch immediately after playing, does the impact happen only if play is immediately before free access to food or does it hang arround?
I no longer need your calculator products
And that is the rub, the main customers left for high end* calculators are students. They buy them because they aren't allowed to take laptops, smartphones etc into exams but in at least some institutions they are allowed to take in particular models of high end calculator.
However high end calculators can also be used as cheating devices. Information that the students are supposed to remember can often be stored as can programs to do things the students are supposed to be able to do manually. Calculator vendors have responsed by introducing "exam modes" where the user is denied access to all memories and with some visual indication to allow invigilators to check the device is in exam mode. However this breaks down if cheaters can replace the firmware with one that pretends to be in exam mode but doesn't actually enforce the restrictions.
My department at uni doesn't allow high end calculators in exams at all . Personally I think that is the way to go as it's perfectly possible to test what needs to be tested without any questions that require a high end calculator to answer. The calculator vendors certainly wouldn't like that endgame though.
* by which I mean anything with more features than a basic "scientific" calculator.
Is that bandwidth is being shifted from one medium to another through the same output device.
Afaict the main difference is in where that bandwidth is from and to.
With traditional TV (whether delivered over cable, sattelite or terrestrial) the bandwidth is used in a broadcast manner. So on each network segment it's only used once per channel no matter how many users tune into that channel.
With ondemand provided by the last mile communications provider the communications provider has full control over where the bandwidth is to.
With internet TV they can ask the provider nicely to locate their servers more locally but without getting into nasty blocking practices there is little they can do to make them do so and managing n sets of third party servers in each location you want content to be streamed from is going to be a lot more overhead than maintaining one server that you control.
"Last mile" in telecoms reffers to the final connection from communications provider infrastructure to the use. In a big city it may be less than a mile in the countryside it may be significantly more.
Having said that while netflix traffic may not be running accross core internet backbones in signficant quantities I suspect it is going a lot further than the "last mile" connection in many if not most cases.
Some believe that there should be an online money transfer mechanism which offers similar levels of anonymity to cash and like cash can't easilly be blocked by governements. Many governements would rather such a system didn't exist and have legislated accordingly (for example with laws requiring money transfer buisnesses to collect details of their customers, freeze accounts etc).
Bitcoin was designed to be highly distributed and very difficult (though probablly not impossible) to subvert. That is it's attraction.
That and they are actually promoted, a great machine is of little use if noone knows it exists.
I have a HP mini 5101 with the high res screen option and I love the thing for giving me a usable screen resoloution (which IMO was the biggest issue with netbooks for general desktop tasks) in a small package but afaict HP never marketed it. So the only people who would have been likely to end up with one are those carefully researching the market for a machine with those charactersitcs.
MS has integrated instruction set emulation into windows before (they certainly did it on alpha and I think they did it on itanium too since the hardware level emulation was appalling and IIRC it was even removed in later versions of the itanium line), there is no reason they couldn't do it again. Intel are claiming they won't but then it's in intels interests to claim that given that they are trying to compete with ARM.
He may have said that but the fact is.
1: PC game vendors have been cracking down hard on the used game market ever since valve showed the industry they could get away with it.
2: Console game vendors have been taking more "baby steps". They still allow resale for the moment but they increasingly include content that can only be registered to one account.
They could do it steam style, ship a copy of the game encrypted on physical media and then force you to tie the key to your account before you can decrypt and install it. That way they could lock things to a purchaser while keeping bandwidth use low. Especially if they don't insist on updates (IIRC steam did insist on updating to the latest version as part of an install from physical media but I may be wrong there).
Afaict microsims are already as small as microSD, making them even smaller will make them very awkward to handle.
Not a big deal for those who live in one country and use one SIM all the time but a very big deal for those who travel and want to swap sims to get a deal that is decent locally.
I was under the impression that pretty much all VGA cables used a coax based design though on the cheaper cables it's probablly not a very high precision one. Certainly the ones i've cut open on old monitors did (I notice that LCDs tend to come with much thinner cables than the CRTs of old did, I guess they get away with a lower quality cable because of the short run length and the fact they are sampling the signal anyway rather than treating it as a continuous analog signal but I have never cut one of those cables open)
A perfect high frequency signal cable would have a core of perfect conductor surrounded by a perfectly dimensioned* insulator enclosed in a perfectly round and perfectly conductive screen. Of course that perfection is unatainable but the more you pay for your cable the closer they can get to those ideals, particularly the tolerances of the various dimensions and the quality of the screen. Raising the size of the cable (the relative sizes of different components of the cable are constrained by impedance requirements but the design as a whole can be scaled pretty freely) also tends to help both by reducing resistive losses and reducing the impact of sizing errors.
There are standards from the bodies that define HDMI. Since HDMI cables aren't a threat to safety or other critical things and afaict most cables do a pretty good job of complying I don't see any reason to make those standards mandatory.
The reason we have overpriced HDMI cables is twofold
1: Some people belive that more expensive always means better and don't get that with a digital cable things tend to either work perfectly or very obviously not work.
2: big electronics stores operate on the principle of being competitive on the big ticket items while overcharging on the extras. I don't particually like this practice but I don't see any other real way they could compete with online only vendors and I wouldn't want physical stores to dissapear.