I just want to know: if ATI prduced some hardware which works much better than NVIDIA, could NVIDIA sue ATI claiming that "they must have spied upon our work to be so good", then peek at their prodect and copy it?
I think this wouldn't happen... Which is the difference?
BTW, I don't see why Microsoft must know what happened. Or better, they must know what happened but not HOW. If some judge decides Viodentia did not steal the source, Microsoft have to trust him that nothing illegal happened. He sais "he did not steal your source, rest assured" and they do.
I am talking about how I think law should work, not how it actually does, though.
France is a culture of peasants. The government is Paternalistic to the extreme. The citizenry accepts that the government will regulate their lives, and attempts at direct independence are futile. This is why France is the laughingstock of Europe, and that is why the French are so grouchy. Pity them and ignore them.
Everyone can think anything she/he wants about other people. Some also say it, this is the problem.
Super-KamioKande didn't establish that neutrinos had mass directly. For that it needed the neutrinos from the nova to arrive spread over a time greater than that in which they departed. The duration of the beam on the nova was estimated about ten seconds, which is almost the same time spread of the revealed neutrinos.
For the oscillations, they are long known, and one of the most simple and exact explainations is that their eigenstate of mass are not the eigenstates of flavour, which in turn means that they have different mass, and at least one is different from zero. Neutrino Oscillation is not a proof for neutrino to have mass. Just a strong hint that they may have
Quantum crypto is untouched by this. It (at least in E97 protocol) relies on quantum entanglement, which cannot be "read". It is a property of a composite system, which cannot be measured if you own just one. You won't even know whether your qubit is entangled or not. (even if you own two, you cannot know, you need lots of pairs to observe entanglement)
Non destructive measuring is not a problem too: measuring a system means not projecting it in the measured state (as many say, nor it is just adding "noise"): that is the "easy" way of thinking. You can couple your system to an "amplifier" which measures and leaves almost the same state. You will simply not know where it was before.
It was theorically possible, it is now practically.
Sorry, I obviously meant it is not proved in-NP-but-not-in-P. It could exist an algorithm which can factorize numbers in polynomial time, we just don't know. And since proving an algorithm exists is often the same as founding it, proving it outside P is very, very difficult.
Quite right in all the point, except one: factoring is polynomial in QComputing, so you go from time 2^N to time C*N^2.
Of course C will be very high in the first times... The hard work is realizing a scalable system, as you say...
Well, I just think it still is a matter of "if". Having graduated on QC, I may say that the quantum computer may com in 10 or 100 years, or never.
Actually no hardware is found which can be used. The 5-qbits computer is a molecule of some sort, which is addressed and controlled by magnetic fields. This is called Nuclear Magnetic Resonance. Expanding the system means finding another molecule, cheap and easy to manipulate (it should not explode in contact with air, and have some necessary properties), which has more different atoms than the previous one. The fields used to address it should be reconfigured as well.
This means that having a 5-qubit computer does not mean we will ever have a 30-qbits one! Not like we can easily imagine a 1024-bit processor, without them being in commerce.
Maybe the QC will break through, but will need a completely new hardware, different from today's one.
People talking on QC as if it existed, or near to come by, are wrong, IMHO.
Quantum crypto is reality instead. The protocol E97, invented by Arthur Ekert has been tested and found working. So the article is not much more than junk.
Quantum computing is not a reality as far as we know.
Few qubits have been realized with a Nuclear Resonance apparatus, but the system is clearly not scalable, it cannot grow to more than 7 qubits (maybe few more, but it must be reimplemented by scratch on a new molecule). The only thing they could factorize is 15... and it took far more than a microsecond.
Quantum Dot, the most scalable quantum bit ever built, has a 10% error rate, 100000 times more than a relatively bad classical system.
Anyway that experiment proved that Shor's algorith is theoretically valid, we just have to find a completely new way to implement a Quantum Computer. Easy, ain't it?
OTOH quantum crypto has been researched, and a working apparatus realized completely secure (in a fundamental manner) communication over few kilometers (~10Km i think), over the lake of geneve. When I say fundamental, I say that they could understand whether someone was listening or not (on a ship...) so no eavesdropping was possible.
Banking has nothing to fear from quantum computing.
Laws neither, it is just someone which said "oh its too fast" without knowing what it really is.
Factoring is not proven NPC nor NP, meaning tht we do not know whether it is really a complex problem, or we just don't know an algorithm to di it fast.
Quantum Computing can help NPC problems with a square root improvement on complexity (which means, the task requiring N steps, requires now sqr(N) ) over the best classical algoritm.
We just even don't know if there exist any problem in the NP class which we cannot solve in polynomial time.
Quantum computing is not proven to be better than classical yet.
I really think that guy's totally wrong on the points. Think about the features of Firefox while considering the inability of IE to tab-browse anything nor block ads, or do simple tasks that now and then you may need. This is not what I think is lack of innovation. Go on and ask yourself if having virtual desktops or multiple ones is so bad. Does proprietary software have so many ideas? Main reason, I think, is beacuse if someone has an idea in OSS which is not able to implemnt, anyone can start from his work and finish it. If someone of the programmers from Microsoft ever had some really good ideas, do you think they could include it in their work freely?
On professionalism, just think about the development of ReiserFS. Company-like development had thoso guys writing by scratch Reiser4 while still on bugfixes of Reiser3. That is real professionalism. NTFS came so lately just to fix the increscious lack of permissions of VFAT...
Conceptual Integrity? might be true. On must admit that any piece of your OS system may have an upgrade at any time, independently. You must decide whether it is good or bad to have your WM upgraded together with your browser, your kernel, shell and various tools each three-or-so yeras or having a brand new version of each one every few months or so. Consider that even when you chose the bundle pack you may always be forced to upgrade or bugfix periodically...
OSS has problems, but are not that ones. IMHO the main one is money. With microsoft spending 100 million $ to promote (I may say some of them went in that author's pocket), OSS must rely on being what one wants, on people trying and speaking with friends, and so on. Must not be so a bad paradigm if it is still alive...
For what has being said, another problem is the lack of contents. OS game engines are more and more often good state of the art code, or so, but still they cant beat Doom3 due to the poorness of the textures. Not any program is well localized, or so. Those things are not funny as coding, and people doesn't do that unless thei're paid. That is a problem for which I have no solutions. Not those that guy pointed out
Slashdot Mirror
I have no doubt that you are right.
I just want to know: if ATI prduced some hardware which works much better than NVIDIA, could NVIDIA sue ATI claiming that "they must have spied upon our work to be so good", then peek at their prodect and copy it?
I think this wouldn't happen... Which is the difference?
BTW, I don't see why Microsoft must know what happened. Or better, they must know what happened but not HOW. If some judge decides Viodentia did not steal the source, Microsoft have to trust him that nothing illegal happened. He sais "he did not steal your source, rest assured" and they do.
I am talking about how I think law should work, not how it actually does, though.
We happen to be only the third most intelligent ones...
They actually link to linux.com.
Please tell me what's "expected", after a full load of bad live demos and delays...
"a few exceptions does not disprove the merit of the generalization."
Really, exceptions disprove the rule, contrary to everything you seem to think.
Everyone can think anything she/he wants about other people. Some also say it, this is the problem.
... having RMS after you mustn't be a nice experience! He is the real victim of this. :P
I would not label the devices, would seem some kind of "standard compliance" advice. People would buy them because they just "work".
Super-KamioKande didn't establish that neutrinos had mass directly. For that it needed the neutrinos from the nova to arrive spread over a time greater than that in which they departed. The duration of the beam on the nova was estimated about ten seconds, which is almost the same time spread of the revealed neutrinos.
For the oscillations, they are long known, and one of the most simple and exact explainations is that their eigenstate of mass are not the eigenstates of flavour, which in turn means that they have different mass, and at least one is different from zero. Neutrino Oscillation is not a proof for neutrino to have mass. Just a strong hint that they may have
Quantum crypto is untouched by this.
It (at least in E97 protocol) relies on quantum entanglement, which cannot be "read". It is a property of a composite system, which cannot be measured if you own just one.
You won't even know whether your qubit is entangled or not.
(even if you own two, you cannot know, you need lots of pairs to observe entanglement)
Non destructive measuring is not a problem too: measuring a system means not projecting it in the measured state (as many say, nor it is just adding "noise"): that is the "easy" way of thinking.
You can couple your system to an "amplifier" which measures and leaves almost the same state. You will simply not know where it was before.
It was theorically possible, it is now practically.
Sorry, I obviously meant it is not proved in-NP-but-not-in-P.
It could exist an algorithm which can factorize numbers in polynomial time, we just don't know. And since proving an algorithm exists is often the same as founding it, proving it outside P is very, very difficult.
Quite right in all the point, except one: factoring is polynomial in QComputing, so you go from time 2^N to time C*N^2. Of course C will be very high in the first times... The hard work is realizing a scalable system, as you say...
Well, I just think it still is a matter of "if".
Having graduated on QC, I may say that the quantum computer may com in 10 or 100 years, or never.
Actually no hardware is found which can be used. The 5-qbits computer is a molecule of some sort, which is addressed and controlled by magnetic fields. This is called Nuclear Magnetic Resonance. Expanding the system means finding another molecule, cheap and easy to manipulate (it should not explode in contact with air, and have some necessary properties), which has more different atoms than the previous one. The fields used to address it should be reconfigured as well.
This means that having a 5-qubit computer does not mean we will ever have a 30-qbits one! Not like we can easily imagine a 1024-bit processor, without them being in commerce.
Maybe the QC will break through, but will need a completely new hardware, different from today's one.
People talking on QC as if it existed, or near to come by, are wrong, IMHO.
Quantum crypto is reality instead. The protocol E97, invented by Arthur Ekert has been tested and found working. So the article is not much more than junk.
Actually, this is far too useless article.
Quantum computing is not a reality as far as we know.
Few qubits have been realized with a Nuclear Resonance apparatus, but the system is clearly not scalable, it cannot grow to more than 7 qubits (maybe few more, but it must be reimplemented by scratch on a new molecule).
The only thing they could factorize is 15... and it took far more than a microsecond.
Quantum Dot, the most scalable quantum bit ever built, has a 10% error rate, 100000 times more than a relatively bad classical system.
Anyway that experiment proved that Shor's algorith is theoretically valid, we just have to find a completely new way to implement a Quantum Computer. Easy, ain't it?
OTOH quantum crypto has been researched, and a working apparatus realized completely secure (in a fundamental manner) communication over few kilometers (~10Km i think), over the lake of geneve.
When I say fundamental, I say that they could understand whether someone was listening or not (on a ship...) so no eavesdropping was possible.
Banking has nothing to fear from quantum computing.
Laws neither, it is just someone which said "oh its too fast" without knowing what it really is.
Factoring is not proven NPC nor NP, meaning tht we do not know whether it is really a complex problem, or we just don't know an algorithm to di it fast.
Quantum Computing can help NPC problems with a square root improvement on complexity (which means, the task requiring N steps, requires now sqr(N) ) over the best classical algoritm.
We just even don't know if there exist any problem in the NP class which we cannot solve in polynomial time.
Quantum computing is not proven to be better than classical yet.
I really think that guy's totally wrong on the points. Think about the features of Firefox while considering the inability of IE to tab-browse anything nor block ads, or do simple tasks that now and then you may need. This is not what I think is lack of innovation. Go on and ask yourself if having virtual desktops or multiple ones is so bad. Does proprietary software have so many ideas? Main reason, I think, is beacuse if someone has an idea in OSS which is not able to implemnt, anyone can start from his work and finish it. If someone of the programmers from Microsoft ever had some really good ideas, do you think they could include it in their work freely? On professionalism, just think about the development of ReiserFS. Company-like development had thoso guys writing by scratch Reiser4 while still on bugfixes of Reiser3. That is real professionalism. NTFS came so lately just to fix the increscious lack of permissions of VFAT... Conceptual Integrity? might be true. On must admit that any piece of your OS system may have an upgrade at any time, independently. You must decide whether it is good or bad to have your WM upgraded together with your browser, your kernel, shell and various tools each three-or-so yeras or having a brand new version of each one every few months or so. Consider that even when you chose the bundle pack you may always be forced to upgrade or bugfix periodically... OSS has problems, but are not that ones. IMHO the main one is money. With microsoft spending 100 million $ to promote (I may say some of them went in that author's pocket), OSS must rely on being what one wants, on people trying and speaking with friends, and so on. Must not be so a bad paradigm if it is still alive... For what has being said, another problem is the lack of contents. OS game engines are more and more often good state of the art code, or so, but still they cant beat Doom3 due to the poorness of the textures. Not any program is well localized, or so. Those things are not funny as coding, and people doesn't do that unless thei're paid. That is a problem for which I have no solutions. Not those that guy pointed out