If the above was even one bit true, non-GPL modules would have been completely forbidden from loading in the kernel at all.
Nothing I said was contrary to what you said.
The derived work of the GPL kernel and non-GPL NVidia module is produced when the module is loaded. This is when the linking of both happens. It is very important to note this, because the "calling" thing you are using is just red-herring.
The whole purpose of linking a module is to enable calling code from that module. So I implied that.
This derived work is allowed because it is done by the user of the system. The result never leaves the memory of the user's system, so it is never distributed. GPL explicitly allows the user to do anything, as long as it doesn't involve distributing the result.
It also means that GPL kernel + non-GPL-nvidia cannot be distributed together.
If you take a look of the code that compiles the NVidia module for your kernel, you will notice that the code does not contain any GPL code in itself. It may require some kernel headers and config files, but these have been established to not be copyrightable (and thus no need of license for them).
Like I said: if you distribute such a (derivative) work, it needs to have a GPL-compatible license and provide sources.
Just using the code has never been seen as distribution, so I didn't mention that.
The problem is that some kernel developers have implemented DRM system, that artificially limits the user in what he can do with his system. If he tries to compile a module that is not under GPL license, a selected number of functions would cause the build to fail. The freedom of the user is artificially taken away. Well, thanks to the GPL the user can hack the kernel and remove the marks of these function, however this is procedure that takes time and effort that are basically wasted.
This is besides the point. The compilation checks are just there so people don't accidentally create derived works by calling internal kernel code (non-public interfaces). It would be trivial to disable those checks, but that wouldn't make it suddenly OK to call the internal kernel code, from a legal point of view.
This is what NVidia wants removed. These functions are not special in any legal way. All kernel functions called by any module are covered under GPL, because the whole kernel (including the files from BSD) is GPL.
This is incorrect. The Linux kernel is GPL, but has exemptions for calling functions/syscalls that are part of the public interface (otherwise all Linux apps would have to be GPLv2). Binary-only drivers that call non-public code are only allowed so long as they are not distributed with the kernel.
There is a legal grey area regarding whether or not dynamic linking GPL code constitutes a derivative work (FSF thinks it does). It would have to be decided by court, but using these internal kernel functions would likely be seen as creating a derivative work, regardless of any compile checks.
That's not what's happening at all.
The basic fact is: the Linux kernel sources are GPL licensed. This was an early decision by Linus, and no amount of wishing will change that. There are just to many contributors that would have to approve a re-license.
Now, the GPL is very clear regarding derived work: if you distribute such a work, it needs to have a GPL-compatible license and provide sources.
What constitutes a derived work for a kernel? Basically, calling any code from the kernel would create a derived work, so the Linux license contains exemptions for user space code that calls the kernel through the public interface.
However, in this case the nVidia driver would call an internal kernel function, that is not exempted, so this would create a GPL derived work. The function is so low level that it would create an intimate bond between the Linux kernel and the nVidia binary driver.
Even if the Linux maintainers would allow this, anyone that wrote any part of Linux could start a court case against nVidia for breach of license. Would that be a better outcome?
In this case it was either because my brain anticipated the 'q' of the word 'qwerty' and skipped ahead, or because the word 'qwerty' temporarily fooled my brain into 'qwerty mode' (I can type in both layouts). It's probably the first one, since I often mistype depending on what I'm thinking of. Sometimes while I'm typing a sentence, someone distracts me for a split second by talking, and then I notice I typed a word or idea related to what was said. The subconscious brain seems to have these independent modules that act like dumb subroutines. They react quickly to input, but are not smart enough to filter out unwanted input.
Just google LayerCake and you will see some press releases from early 2012 where they announce they can run ARM code. I can't find any technical specs on BlueStacks, so it's all marketing speak.
Again: most games use ARM native code, so Android x86 would not be enough to run that. You still need an ARM emulator/translator.
As it happens, Intel created an ARM translation library (Houdini) for their Medfield phones. It allows the Medfield x86 Android to run ARM code. Some resourceful people managed to integrate this lib in the desktop x86 Android: link. It's possible BlueStacks does the same thing, or something similar.
Actually, while technical details are sparse, Bluestacks seems to emulate ARM instructions with their LayerCake technology, which also provides hardware gfx acceleration. Most Android games use native ARM code in a Dalvik/java wrapper, and Bluestacks seems to be able to run at least some unmodified.
What'd they do, shift all the load to AMD servers?
Nah, they found an old Pentium 4 desktop, switched that on, and ran Crysis on it. But they had to lower the gfx settings because all the lights started to dim.
Scientist 1 : Look! We found these crystals with dots on it. We believe they're some ancient data storage discs.
Scientist 2 : Cool! What do they say?
Scientist 1 : We don't know, we need the software to decode them.
Scientist 2 : And where is the software?
Scientist 1 : We're pretty sure it's on one of the discs...
(Scientist 1 : Also, we need a running DRM server, whatever that may be)
As far as psychologically addictive, there is no such thing. Addiction is bio-chemical. You cannot be "addicted" to gambling, shopping, masturbation, etc. These are compulsive behaviors - they are NOT addiction. That is not meant to insult sufferers, compulsive behavior patterns are an illness and are FAR worse to deal with than simple physical addition. With physical addiction, you simply need to detox.
Recent research seems to indicate that the brain reacts the same to psychological addiction as to substance addiction. So both seem to be physically addictive.
Addiction researchers have turned to neuroscience to understand the mechanisms underlying different addictions. But the more they have tried to tease apart substance and behavioural addiction, the more similarities they find. For one thing, the brains of people with addiction look similar whether the addiction is to a substance or to a behaviour. When the brains of addicts are scanned using functional magnetic resonance imaging (fMRI) while they are shown videos of the addictive behaviour or of people using their drug of choice, the same reward centres are activated, says Robert Malenka, a Stanford neuroscientist who studies addiction.
Well, one is a compound containing an oxygen atom bonded to two hydrocarbon groups, and the other is a field that permeates all of space-time.
But you probably meant aether;-).
Aether was conceived as a substance that permeated 3-space. Light would travel through it, like a wave through water. It would cause the speed of light in vacuum to be variable, depending on one's velocity, which is not what was observed.
The Higgs field seems to be a property of 4-space (space-time), or a field that permeates it (we're not sure at this point what it is exactly). Everything (including light) travels through 4-space, and the speed of light in vacuum measures the same in all reference frames.
Personally, I think space-time is a 4-dimensional fluid, which we perceive as continuous, but which at the smallest scale consists of units that shift arrangements and interconnections (like molecules in water, but in 4D). Particles would be just waves in this fluid (like for example a vortex ring in 3D fluid). One such wave could even be observed as different particles (a bit like seeing a cloud and either recognizing a dog or a bunny). The Higgs field could then be a fundamental property of this fluid (something like viscosity of a 3D fluid), so it would affect both the waves/particles that move through it, and the fluid itself.
navigation prompts, restuarant reviews, contact information, descriptions of art, and much more
While it has the potential to do all this, its primary function will be to research what we look at, and serve ads to match. If you think the billboards and neons are bad already, just wait until you put on one of these...
It's not hero worship, it's a matter of respect. Turing made great contributions to several fields related to computers and math. As a reward, he was prosecuted by his own government. So the least we can do is remember his name correctly.
And "honor" is the US spelling of the UK "honour". Your reaction is unproductive, childish and cowardly.
Y'know what's even harder to design? Analog computing. Holy cow. Remember, digital computing was invented by Touring before we even had built a computer. It's easy to visualize how it works. My brain explodes though trying to imagine a fuzzy-logic analog equivalent of a touring machine.
His name was Alan Turing. Honor him by at least spelling his name correctly...
You're right, of course. I was sort of simplifying for the international crowd, who might have heard of KUL. There is so much collaboration between Group T and KUL, sharing facilities and professors, working together on the same projects, that Group T can be regarded by the outside world as a part of KUL. At least in spirit, that is. The administration is still mostly independent, and will probably stay that way for quite some time.
Well, the story already said that 0.1bps was possible, as your paper confirms.
But they also postulated that 100x faster speeds would be possible by modulating the neutrino beam. I was just questioning the feasibility of that.
Assume we would use your mechanism of switching the extraction kicker magnet on/off, and thereby creating a sort of binary (on/off) modulation. I suspect the magnet coil can only be pulsed so fast, its reluctance would resist fast changes, an any hysteresis would cause delays in the magnetic field change.
I also still think that detecting and analyzing the neutrinos would take time (at least, more than detecting photons from a fiber channel would), adding further delay to the communication path.
Of course we won't be sure until someone actually tries it, so if some suckers^H^H^H^H^H^H^Hbankers want to finance it, I'm all for it.;)
Lol. For a moment, I thought you were serious. I don't see why quantum entanglement would be more expensive and lower bandwidth though. This could be done with simple, mass produced consumer electronics/optics. Too bad it doesn't allow actual information transfer.
With the right modulation scheme, this could be increased by at least one or two orders of magnitude
Sure. If anyone ever figures out how to modulate a neutrino beam. Now we're having problems already just detecting them.
I suppose it would be theoretically possible to modulate the accelerator beams, but with the energies involved, I suspect that will be quite a challenge too.
And then there is the small matter of latency. The neutrino beam may be fast, but the whole process of converting the information stream in a modulated beam, and then analyzing the data at the detectors to find the few neutrinos that didn't come from the sun, will take probably more time than sending the data through fiber optics would.
Lol. Non-locality is exactly what makes the "magical spooky action at a distance" possible.
Non-local hidden variables are a completely different beast, compared to classical (local) hidden variables. In fact, some theories are mathematical equivalent to entanglement.
Slashdot Mirror
If the above was even one bit true, non-GPL modules would have been completely forbidden from loading in the kernel at all.
Nothing I said was contrary to what you said.
The derived work of the GPL kernel and non-GPL NVidia module is produced when the module is loaded. This is when the linking of both happens. It is very important to note this, because the "calling" thing you are using is just red-herring.
The whole purpose of linking a module is to enable calling code from that module. So I implied that.
This derived work is allowed because it is done by the user of the system. The result never leaves the memory of the user's system, so it is never distributed. GPL explicitly allows the user to do anything, as long as it doesn't involve distributing the result.
It also means that GPL kernel + non-GPL-nvidia cannot be distributed together.
If you take a look of the code that compiles the NVidia module for your kernel, you will notice that the code does not contain any GPL code in itself. It may require some kernel headers and config files, but these have been established to not be copyrightable (and thus no need of license for them).
Like I said: if you distribute such a (derivative) work, it needs to have a GPL-compatible license and provide sources.
Just using the code has never been seen as distribution, so I didn't mention that.
The problem is that some kernel developers have implemented DRM system, that artificially limits the user in what he can do with his system. If he tries to compile a module that is not under GPL license, a selected number of functions would cause the build to fail. The freedom of the user is artificially taken away. Well, thanks to the GPL the user can hack the kernel and remove the marks of these function, however this is procedure that takes time and effort that are basically wasted.
This is besides the point. The compilation checks are just there so people don't accidentally create derived works by calling internal kernel code (non-public interfaces). It would be trivial to disable those checks, but that wouldn't make it suddenly OK to call the internal kernel code, from a legal point of view.
This is what NVidia wants removed. These functions are not special in any legal way. All kernel functions called by any module are covered under GPL, because the whole kernel (including the files from BSD) is GPL.
This is incorrect. The Linux kernel is GPL, but has exemptions for calling functions/syscalls that are part of the public interface (otherwise all Linux apps would have to be GPLv2). Binary-only drivers that call non-public code are only allowed so long as they are not distributed with the kernel.
There is a legal grey area regarding whether or not dynamic linking GPL code constitutes a derivative work (FSF thinks it does). It would have to be decided by court, but using these internal kernel functions would likely be seen as creating a derivative work, regardless of any compile checks.
APIs GPL only? Seriously guys, WHAT THE FUCK?
That's not what's happening at all.
The basic fact is: the Linux kernel sources are GPL licensed. This was an early decision by Linus, and no amount of wishing will change that. There are just to many contributors that would have to approve a re-license.
Now, the GPL is very clear regarding derived work: if you distribute such a work, it needs to have a GPL-compatible license and provide sources.
What constitutes a derived work for a kernel? Basically, calling any code from the kernel would create a derived work, so the Linux license contains exemptions for user space code that calls the kernel through the public interface.
However, in this case the nVidia driver would call an internal kernel function, that is not exempted, so this would create a GPL derived work. The function is so low level that it would create an intimate bond between the Linux kernel and the nVidia binary driver.
Even if the Linux maintainers would allow this, anyone that wrote any part of Linux could start a court case against nVidia for breach of license. Would that be a better outcome?
In this case it was either because my brain anticipated the 'q' of the word 'qwerty' and skipped ahead, or because the word 'qwerty' temporarily fooled my brain into 'qwerty mode' (I can type in both layouts).
It's probably the first one, since I often mistype depending on what I'm thinking of. Sometimes while I'm typing a sentence, someone distracts me for a split second by talking, and then I notice I typed a word or idea related to what was said.
The subconscious brain seems to have these independent modules that act like dumb subroutines. They react quickly to input, but are not smart enough to filter out unwanted input.
I particularly liked the part when they got stuck in a giant geode.
That was actually the more plausible part of the movie.
Just google LayerCake and you will see some press releases from early 2012 where they announce they can run ARM code. I can't find any technical specs on BlueStacks, so it's all marketing speak.
Again: most games use ARM native code, so Android x86 would not be enough to run that. You still need an ARM emulator/translator.
As it happens, Intel created an ARM translation library (Houdini) for their Medfield phones. It allows the Medfield x86 Android to run ARM code. Some resourceful people managed to integrate this lib in the desktop x86 Android: link. It's possible BlueStacks does the same thing, or something similar.
Actually, while technical details are sparse, Bluestacks seems to emulate ARM instructions with their LayerCake technology, which also provides hardware gfx acceleration.
Most Android games use native ARM code in a Dalvik/java wrapper, and Bluestacks seems to be able to run at least some unmodified.
on q qwerty keyboard...
Lol. Just thinking of qwerty made me type a Q instead of an A (azerty keyboard).
It's most probably a typo of the word "Inquiry". The keys R and T are adjacent on q qwerty keyboard...
What'd they do, shift all the load to AMD servers?
Nah, they found an old Pentium 4 desktop, switched that on, and ran Crysis on it.
But they had to lower the gfx settings because all the lights started to dim.
Scientist 1 : Look! We found these crystals with dots on it. We believe they're some ancient data storage discs.
Scientist 2 : Cool! What do they say?
Scientist 1 : We don't know, we need the software to decode them.
Scientist 2 : And where is the software?
Scientist 1 : We're pretty sure it's on one of the discs...
(Scientist 1 : Also, we need a running DRM server, whatever that may be)
The Libre documentation PDF is awesome and extensive (like 400 pages!) but mentions no word of Pivot Tables.
The top google search result for [libreoffice pivot table] (disclaimer: I never used LibreOffice Calc):
http://help.libreoffice.org/Calc/Pivot_Table
As far as psychologically addictive, there is no such thing. Addiction is bio-chemical. You cannot be "addicted" to gambling, shopping, masturbation, etc. These are compulsive behaviors - they are NOT addiction. That is not meant to insult sufferers, compulsive behavior patterns are an illness and are FAR worse to deal with than simple physical addition. With physical addiction, you simply need to detox.
Recent research seems to indicate that the brain reacts the same to psychological addiction as to substance addiction. So both seem to be physically addictive.
From http://www.newscientist.com/article/mg21528810.200-the-people-who-are-addicted-to-addiction.html:
Addiction researchers have turned to neuroscience to understand the mechanisms underlying different addictions. But the more they have tried to tease apart substance and behavioural addiction, the more similarities they find. For one thing, the brains of people with addiction look similar whether the addiction is to a substance or to a behaviour. When the brains of addicts are scanned using functional magnetic resonance imaging (fMRI) while they are shown videos of the addictive behaviour or of people using their drug of choice, the same reward centres are activated, says Robert Malenka, a Stanford neuroscientist who studies addiction.
Well, one is a compound containing an oxygen atom bonded to two hydrocarbon groups, and the other is a field that permeates all of space-time. ;-) .
But you probably meant aether
Aether was conceived as a substance that permeated 3-space. Light would travel through it, like a wave through water. It would cause the speed of light in vacuum to be variable, depending on one's velocity, which is not what was observed.
The Higgs field seems to be a property of 4-space (space-time), or a field that permeates it (we're not sure at this point what it is exactly). Everything (including light) travels through 4-space, and the speed of light in vacuum measures the same in all reference frames.
Personally, I think space-time is a 4-dimensional fluid, which we perceive as continuous, but which at the smallest scale consists of units that shift arrangements and interconnections (like molecules in water, but in 4D). Particles would be just waves in this fluid (like for example a vortex ring in 3D fluid). One such wave could even be observed as different particles (a bit like seeing a cloud and either recognizing a dog or a bunny). The Higgs field could then be a fundamental property of this fluid (something like viscosity of a 3D fluid), so it would affect both the waves/particles that move through it, and the fluid itself.
navigation prompts, restuarant reviews, contact information, descriptions of art, and much more
While it has the potential to do all this, its primary function will be to research what we look at, and serve ads to match.
If you think the billboards and neons are bad already, just wait until you put on one of these...
It's not hero worship, it's a matter of respect. Turing made great contributions to several fields related to computers and math. As a reward, he was prosecuted by his own government. So the least we can do is remember his name correctly.
And "honor" is the US spelling of the UK "honour". Your reaction is unproductive, childish and cowardly.
Y'know what's even harder to design? Analog computing. Holy cow. Remember, digital computing was invented by Touring before we even had built a computer. It's easy to visualize how it works. My brain explodes though trying to imagine a fuzzy-logic analog equivalent of a touring machine.
His name was Alan Turing. Honor him by at least spelling his name correctly...
China even already has a strain of red algae in the mountains. Though it's probably not that one, since it only grows in one (large) area.
You're right, of course. I was sort of simplifying for the international crowd, who might have heard of KUL. There is so much collaboration between Group T and KUL, sharing facilities and professors, working together on the same projects, that Group T can be regarded by the outside world as a part of KUL. At least in spirit, that is. The administration is still mostly independent, and will probably stay that way for quite some time.
the efforts of a 16-person team called "Group T"
Actually Group T is the name of an engineering division of KU Leuven (T=technical). This makes it sound like they made up some team name.
Did they remember to attach a shark to it? It can't possibly function without one!
One mebibyte was just 1/1024 of a gibibyte!
FTFY
FTFY
FTFY
Well, the story already said that 0.1bps was possible, as your paper confirms.
;)
But they also postulated that 100x faster speeds would be possible by modulating the neutrino beam. I was just questioning the feasibility of that.
Assume we would use your mechanism of switching the extraction kicker magnet on/off, and thereby creating a sort of binary (on/off) modulation. I suspect the magnet coil can only be pulsed so fast, its reluctance would resist fast changes, an any hysteresis would cause delays in the magnetic field change.
I also still think that detecting and analyzing the neutrinos would take time (at least, more than detecting photons from a fiber channel would), adding further delay to the communication path.
Of course we won't be sure until someone actually tries it, so if some suckers^H^H^H^H^H^H^Hbankers want to finance it, I'm all for it.
Lol. For a moment, I thought you were serious.
I don't see why quantum entanglement would be more expensive and lower bandwidth though. This could be done with simple, mass produced consumer electronics/optics. Too bad it doesn't allow actual information transfer.
With the right modulation scheme, this could be increased by at least one or two orders of magnitude
Sure. If anyone ever figures out how to modulate a neutrino beam. Now we're having problems already just detecting them.
I suppose it would be theoretically possible to modulate the accelerator beams, but with the energies involved, I suspect that will be quite a challenge too.
And then there is the small matter of latency. The neutrino beam may be fast, but the whole process of converting the information stream in a modulated beam, and then analyzing the data at the detectors to find the few neutrinos that didn't come from the sun, will take probably more time than sending the data through fiber optics would.
Lol. Non-locality is exactly what makes the "magical spooky action at a distance" possible.
Non-local hidden variables are a completely different beast, compared to classical (local) hidden variables. In fact, some theories are mathematical equivalent to entanglement.