No. Try reading the rest of the thread. A modern GPU is as you describe. The 2004 GPUs that we are talking about used fixed function hardware to do these things. They were not programmable vector processors.
While it would apply to written reproduction of that plan, it would not cover the stunt itself. Copyright is the right to control copying of the written word. Although it has been extended from the written word into other media (for example photographs), it does not yet extend to stunts.
Last time that I checked Coca-Cola do sell their product in the uk. Here is a description of UK law. So as I already said it is quite insane to pass laws to give protection to "secrets". The best way to protect them is to ensure that they remain secret, not having laws to punish people who leak them.
No, what the GP said was that essentials should be free. This is not the same as classical communism. As you've said in classical communism everyone gets what they need (so that part is the same) and contributes what they can. The second part is not necessarily what the GP said, although it is one possible interpretation.
The other interpretation is that everyone contributes what they must. And this leave no restriction on the stuff left over (that does not form the basic necessities). As far as I know that is not an economic system that has been tried, although some of the socialist states in Europe approach it.
To be clear: what the GP suggested could be achieved by a dual-economic system. A communist model at the bottom for those goods deemed to be necessities, with another system (e.g a capitalist free market) over the remaining "luxuries". I'm not suggesting that it would work, although I think it would be interesting to see exactly how it did function.
Take a look at the explanation part of the video. The background texture is tiled. You can see some strange deformation in the regular pattern where the object used to be. Also in the drain example there is a strange crater effect as the camera angle changes.
It seems like smooth colour graduations work well, but patterned backgrounds have more obvious deformations.
So under American law: anything that a company attempts to keep secret for commercial benefit has federal protection, and it's a crime to spread that information? What a crazy country you guys live in. That is bizarre.
But trade secrets don't have any actual protection under law though - that was the original point of patents. Surely to qualify as a trade secret he would have needed a non-disclosure agreement, and that contract would have offered him the legal protection anyway?
If only there were some class of goods that people desired, but which were not necessary. Then the market that you describe could co-exist with a world in which people were not homeless, starving or dying of thirst.
It is arguable that NEP was similar to such a system, although it is not obvious that they are the same. One important difference is that the percentage of the economy devoted to producing necessities is much smaller in a modern industrial society than it was in Russia immediately after the war.
Playing Devil's Advocate for a second: why shouldn't they? When did we grant Intellectual Property rights to plans for promotional stunts. How exactly does he feel that he has been violated - copyright (not applicable), patent (no applicable).... Unless he got them to sign some sort of contract before showing them the plans he has no protection..... and now I'll probably RTFA to discover which of these was true.
No, my argument doesn't claim that Microsoft's patent is valid for anything. At no point have I claimed it is a good patent or describes a technique that should get patent protection. And yes on today's general purpose hardware the use is completely obvious.
I have only claimed one thing: that it was not obvious that the hardware of 2004 would accelerate video compression. In fact, I'll explain the problem with a quote:
It simply didn't use to speed up video compression much.
Wow.... I'd never heard of that board and it sounds way ahead of its time. That is pretty amazing for the early 90s. My first accelerator was an nVidia-256 several years later and it had nothing like that level of programability. Why on earth were they not more popular / widely remembered?
So you're arguing that it is obvious in 2004 that a GPU can do video compression because lots of other parts (that the GPU doesn't include) can be combined into something that does video compression..... Yeah, I'm the one missing the clue. Here is a bicycle. It includes none of the parts necessary to encode video, but they all exist in other products. It must be "obvious" that bicycles will encode video.
Fair enough point, really. I've been roundly jumped on by a million people in this thread so it is time for me to concede. Personally I wouldn't have thought it obvious programming GPUs in 2004. It seems clear that you would have done.
Haha. Did you end up at Clearspeed? They would have been called Pixel Fusion back then wouldn't they?
The difference is that you were working for a hardware company and there was more flexibility in how to fit the instruction set to the application. The graphics card market in 2004 wasn't yet going down the same route. Pixel-shaders were still made of blocks of fixed-function instructions rather than a more general model. The programming interface was still through rendering steps / with some invisible compiler transformations.
So while ATi and nVidia may have been looking in the same direction planning to make things more flexible, Microsoft must have written this patent as users, rather than designers, of GPUs. It's a bad patent for other reasons but at that stage it was not obvious that video encoding would be a good match for a pixel shader.
Yes. Moving from the decompression to the encoding was a large difference. At that point in time these cards were not vector processors to simply offload code onto. There was no control-flow (that was a pretty big difference on the Nvidia 6-series and very badly implemented).
Every algorithm needed a different bag of tricks to get it working on GPU hardware, and depending on the exact details sometimes they would speed up, and sometimes they would slow down. Simply being in the same rough area (doing some vector stufF) does not mean that they would respond to the GPU in the same way.
In 2004 it was not obvious at all that the pixel shaders available would not provide enough power or flexibility to handle video encoding. This is not support for the patent - I've read it and it shouldn't have been granted - but the original point made that it is obvious that a graphic card should accelerate video is simply wrong.
No, but providing someone with a unit to calculate (for example) a transformation doesn't mean that you give them an efficient way of computing FFTs. Of course if you give them a general purpose matrix-vector multiplier then it does. In 2004 a graphics card used a fixed function unit, and today it uses a general purpose one. Assumptions about how obvious it is that other applications can be performed don't carry back to previous generations of the hardware.
No, I'm not being sarcastic and I'm not trolling. I'm just better informed than you are. The hardware available in 2004 was not a general purpose vector accelerator. I know because that is what I was doing in 2004 - writing GPGPU code in OpenGL. What ATI did within a year was add specialised hardware to perform video decoding. This is not the same thing and cannot be reused for the encoding tasks that Microsoft are patenting.
So you, the OP and the other reply (not to mention the mods) are all wrong BECAUSE YOU ARE ASSUMING IT IS OBVIOUS NOW GIVEN OUR PROGRAMMABLE VECTOR ARRAYS. All of you lack the experience or the memory to say how obvious it would be ON THE HARDWARE AVAILABLE THEN.
No. You are confusing a current graphics card for one of six years ago. It may be obvious now that we can do fast vector processing on a GPU, but they have come along way in six years, each generation aiming more at GPGPU.
In 2004 the 6600 was just making it onto reviewers desks. CUDA was a gleam in some engineers eye and most graphics cards still used fixed function pipelines to perform rendering. Even the latest generation that allowed programable pixel shaders did not automatically accelerate any vectorisable problem. Scatter was impossible to perform just using rendering steps and gather was expensive.
So again I'll ask you; why is it obvious that units designed to do transformation, lighting and rasterisation (not general vector operations) should be good at encoding video?
Why would it be obvious that hardware designed to accelerate 3d rendering - transformation, lighting and rasterisation - can accelerate the compression of video frames?
Again you seem to be deliberately confusing things. Raw intelligence correlates with successful results in education. Use of birth control correlates with the amount of eduction provided. They are not the same.
It's nice that you have this pet theory that use of birth control will be selected away within a few generations. If you had some kind of evidence it would be interesting. But the "logic" that you are trying to use to argue your case is flawed.
The part of the sentence that you didn't quote provided the meaning for "deliberately". Why do you think that selective quoting allows you to make a strawman?
Slashdot Mirror
No. Try reading the rest of the thread. A modern GPU is as you describe. The 2004 GPUs that we are talking about used fixed function hardware to do these things. They were not programmable vector processors.
While it would apply to written reproduction of that plan, it would not cover the stunt itself. Copyright is the right to control copying of the written word. Although it has been extended from the written word into other media (for example photographs), it does not yet extend to stunts.
Last time that I checked Coca-Cola do sell their product in the uk. Here is a description of UK law. So as I already said it is quite insane to pass laws to give protection to "secrets". The best way to protect them is to ensure that they remain secret, not having laws to punish people who leak them.
No, what the GP said was that essentials should be free. This is not the same as classical communism. As you've said in classical communism everyone gets what they need (so that part is the same) and contributes what they can. The second part is not necessarily what the GP said, although it is one possible interpretation.
The other interpretation is that everyone contributes what they must. And this leave no restriction on the stuff left over (that does not form the basic necessities). As far as I know that is not an economic system that has been tried, although some of the socialist states in Europe approach it.
To be clear: what the GP suggested could be achieved by a dual-economic system. A communist model at the bottom for those goods deemed to be necessities, with another system (e.g a capitalist free market) over the remaining "luxuries". I'm not suggesting that it would work, although I think it would be interesting to see exactly how it did function.
Take a look at the explanation part of the video. The background texture is tiled. You can see some strange deformation in the regular pattern where the object used to be. Also in the drain example there is a strange crater effect as the camera angle changes.
It seems like smooth colour graduations work well, but patterned backgrounds have more obvious deformations.
So under American law: anything that a company attempts to keep secret for commercial benefit has federal protection, and it's a crime to spread that information? What a crazy country you guys live in. That is bizarre.
But trade secrets don't have any actual protection under law though - that was the original point of patents. Surely to qualify as a trade secret he would have needed a non-disclosure agreement, and that contract would have offered him the legal protection anyway?
If only there were some class of goods that people desired, but which were not necessary. Then the market that you describe could co-exist with a world in which people were not homeless, starving or dying of thirst.
It is arguable that NEP was similar to such a system, although it is not obvious that they are the same. One important difference is that the percentage of the economy devoted to producing necessities is much smaller in a modern industrial society than it was in Russia immediately after the war.
Playing Devil's Advocate for a second: why shouldn't they? When did we grant Intellectual Property rights to plans for promotional stunts. How exactly does he feel that he has been violated - copyright (not applicable), patent (no applicable).... Unless he got them to sign some sort of contract before showing them the plans he has no protection..... and now I'll probably RTFA to discover which of these was true.
Where are these GPUs that have general purpose computing capabilities that have been around for decades?
No, my argument doesn't claim that Microsoft's patent is valid for anything. At no point have I claimed it is a good patent or describes a technique that should get patent protection. And yes on today's general purpose hardware the use is completely obvious.
I have only claimed one thing: that it was not obvious that the hardware of 2004 would accelerate video compression. In fact, I'll explain the problem with a quote:
Oooooo! Stinging combe-back. Did you get it from a video you watched on your bicycle?
That was fixed functionality hardware, it was not programmable and it could not be accessed from pixel shaders.
Wow.... I'd never heard of that board and it sounds way ahead of its time. That is pretty amazing for the early 90s. My first accelerator was an nVidia-256 several years later and it had nothing like that level of programability. Why on earth were they not more popular / widely remembered?
So you're arguing that it is obvious in 2004 that a GPU can do video compression because lots of other parts (that the GPU doesn't include) can be combined into something that does video compression..... Yeah, I'm the one missing the clue. Here is a bicycle. It includes none of the parts necessary to encode video, but they all exist in other products. It must be "obvious" that bicycles will encode video.
Fair enough point, really. I've been roundly jumped on by a million people in this thread so it is time for me to concede. Personally I wouldn't have thought it obvious programming GPUs in 2004. It seems clear that you would have done.
Haha. Did you end up at Clearspeed? They would have been called Pixel Fusion back then wouldn't they?
The difference is that you were working for a hardware company and there was more flexibility in how to fit the instruction set to the application. The graphics card market in 2004 wasn't yet going down the same route. Pixel-shaders were still made of blocks of fixed-function instructions rather than a more general model. The programming interface was still through rendering steps / with some invisible compiler transformations.
So while ATi and nVidia may have been looking in the same direction planning to make things more flexible, Microsoft must have written this patent as users, rather than designers, of GPUs. It's a bad patent for other reasons but at that stage it was not obvious that video encoding would be a good match for a pixel shader.
Yes. Moving from the decompression to the encoding was a large difference. At that point in time these cards were not vector processors to simply offload code onto. There was no control-flow (that was a pretty big difference on the Nvidia 6-series and very badly implemented).
Every algorithm needed a different bag of tricks to get it working on GPU hardware, and depending on the exact details sometimes they would speed up, and sometimes they would slow down. Simply being in the same rough area (doing some vector stufF) does not mean that they would respond to the GPU in the same way.
In 2004 it was not obvious at all that the pixel shaders available would not provide enough power or flexibility to handle video encoding. This is not support for the patent - I've read it and it shouldn't have been granted - but the original point made that it is obvious that a graphic card should accelerate video is simply wrong.
No, but in 2004 a GPU was not a vector processor. As I explained above.
No, but providing someone with a unit to calculate (for example) a transformation doesn't mean that you give them an efficient way of computing FFTs. Of course if you give them a general purpose matrix-vector multiplier then it does. In 2004 a graphics card used a fixed function unit, and today it uses a general purpose one. Assumptions about how obvious it is that other applications can be performed don't carry back to previous generations of the hardware.
No, I'm not being sarcastic and I'm not trolling. I'm just better informed than you are. The hardware available in 2004 was not a general purpose vector accelerator. I know because that is what I was doing in 2004 - writing GPGPU code in OpenGL. What ATI did within a year was add specialised hardware to perform video decoding. This is not the same thing and cannot be reused for the encoding tasks that Microsoft are patenting.
So you, the OP and the other reply (not to mention the mods) are all wrong BECAUSE YOU ARE ASSUMING IT IS OBVIOUS NOW GIVEN OUR PROGRAMMABLE VECTOR ARRAYS. All of you lack the experience or the memory to say how obvious it would be ON THE HARDWARE AVAILABLE THEN.
No. You are confusing a current graphics card for one of six years ago. It may be obvious now that we can do fast vector processing on a GPU, but they have come along way in six years, each generation aiming more at GPGPU.
In 2004 the 6600 was just making it onto reviewers desks. CUDA was a gleam in some engineers eye and most graphics cards still used fixed function pipelines to perform rendering. Even the latest generation that allowed programable pixel shaders did not automatically accelerate any vectorisable problem. Scatter was impossible to perform just using rendering steps and gather was expensive.
So again I'll ask you; why is it obvious that units designed to do transformation, lighting and rasterisation (not general vector operations) should be good at encoding video?
Why would it be obvious that hardware designed to accelerate 3d rendering - transformation, lighting and rasterisation - can accelerate the compression of video frames?
It seems that you are 'obviously' wrong.
Again you seem to be deliberately confusing things. Raw intelligence correlates with successful results in education. Use of birth control correlates with the amount of eduction provided. They are not the same.
It's nice that you have this pet theory that use of birth control will be selected away within a few generations. If you had some kind of evidence it would be interesting. But the "logic" that you are trying to use to argue your case is flawed.
The part of the sentence that you didn't quote provided the meaning for "deliberately". Why do you think that selective quoting allows you to make a strawman?