Sure if you tell me why Ford would spend a billion to build a car factory even though GM could too.
Yet another pointless rant. Why does Ford spend the billion? Well, because they are the ones who will use it, they don't have to be afraid that GM will take over. Its their building, after all. It is, however, easy to simply rip off a novel technology if it is published.
Actually, IBM and Intell sued for billions over that and lost badly in the courts. So did the PC industry flop? no it boomed, no, in fact after that the IBM compatable PC took over the marketplace and there was a nuclear explosion in R&D, along with a nuclear explosion in business and commerce.
Yes, everyone gained from it, except IBM and Intel, who LOST. In the long term, they won too, but a scenario like this can easily break a company's neck. IBM was (and is) a big player, but imagine the PC had been invented by a small start-up company. They would have gone bankrupt.
The negros shouldn't need their "kind masters" permission to exercise freedom.
So your point is that scientists shall do their expensive and time-consuming R&D for free and not get a single cent paid for it, is that right? You do know that only government-funded research can afford this, right?
Oh, and cut the US crap already. I am not a US citizen.
Invention and creation happened 2000 years before patent, and got lost, forgotten, hidden. The antikythera knowledge was lost, and had to be re-invented, which wasn't guaranteed to happen. In fact, lost inventions may NEVER appear again. Imagine Goedel had never published his work, for example.
The free copy nature of the Internet is something COMPLETELY different. I am talking about inventions, not information. You are mixing patents with copyright.
Also, your problem seems to be that you don't consider how R&D is done in other departments. Equipment costs A LOT, lab costs are high, etc. You don't need that expensive equipment for writing software, you don't need a lab. Guess why FOSS didn't arise for hardware, or pharmaceuticals: BECAUSE R&D IS EXPENSIVE THERE. Getting rid of patents works only when R&D is comparatively cheap and a temporary monopoly is not plausible - which is very true for software, with the addition that major breakthroughs are very rare with software. However, pharma R&D is usually expensive. It *has* to pay off to do some research; how would it pay off if others can simply rip off my hard work? The only alternative would be to lock down the product so no one can reverse-engineer it; of course this also means that the company does not disclose the actual invention. If the company vanishes, the invention vanishes too. It already happened with stuff like Rotodyne.
So, now answer me this question: if R&D is (very) expensive, and everyone is free to rip the results, why should I do research in the first place? It just doesn't pay off!
Also, your comment about Africa shows that you simply didnt read what I said. Again: I AM ALL FOR FAIR USE CLAUSES. Africa should have gotten a permission in the first place.
Most of your points are valid only if you manually install the components. Distros like Ubuntu do take care of the 300MB libraries etc. Some others:
- Today RPM and DEB are the de-facto standard package formats. There are also GUIs for installing these packages, also handling dependencies. (gdebi for DEBs, I dont know how the RPM one is named.) - GStreamer is something like VFW, and is gaining popularity. (Ubuntu uses it by default.) - The directory hierarchy is irrelevant for the home user. Again, take Ubuntu as example. Who cares about where the executables are? Click on the menu, and thats it. Also, its not hard to remember that/usr/bin contains distro stuff and/usr/local/bin manually installed things. (local should be renamed, though.) - The SuSE issue is definitely a design error. Note however that Windows has similar issues as well. - Loading the libraries is slow, yes, partially because of missing symbol visibility (all symbols are loaded). KDE will make use of the gcc visibility feature to address this. As for Firefox, I don't know. - *Advanced* administration is best done via command-line, because it has much greater flexibility than a GUI could ever have. Also, it scales better. Now, a list of installed codecs is available via gstreamer IIRC. Also, "bum" is a tool for administering the programs loaded at startup. It needs to be more known, though. - I agree that there seems to be a bloat-mania in Linux land. Foobar2000 is *very* well coded. MPC, however, uses the VFW codecs, so the comparison with mplayer isnt really fair. vlc, mplayer, and xine seem to link everything, though (mplayer.exe is 9MB monster in Windows). Still, if I want to play *any* video, I use vlc, precisely because it has support for bazillion formats (even ones not supported in Foobar).
In sum, you are partially right, but as I said, most points are really a distro issue. Also keep in mind that most people do NOT install Windows by themselves; instead, they use what's on that Gateway PC they just bought. When Windows is borked again and infected with 15911 viruses after 6 months, Generic Geek is called and he sets up a fresh one. Or people call support, and get told to use the Recovery CD. With Linux it would be the same (it would be probably more robust, though).
So, the REAL reason why Linux isn't there yet is because it needs to be preinstalled on retail PCs.
"massive restrictions on what people can copy" - Without patents, people could not copy anything BECAUSE THERE WOULD BE NOTHING TO COPY FROM. The science behind the drugs would be locked away deep in the company HQs. And no, people would NOT "invent them anyhow", precisely because the research is incremental. Company A makes small steps, patent them, B licenses A' patents, makes small steps... your alternative would result in stagnation because no one would be willing to disclose their results - everything would be secret. With patents, the results are PUBLIC.
Also, with patents, countries have the option to ignore them, like Brazil did. Keep in mind that some African countries are considering this as well. This is why I say that the enforcement is broken: there is no "fair use" clause for patents, save for usage by the military. THIS is what has to be changed. Africa is obviously in a state of emergency and should get free access to those patents.
As stated before, the AIDS problem is caused by a broken patent ENFORCEMENT. Without patents, the Africans *still* would not get any generics, except that the science behind them would be kept secret.
Blame the insane intellectual property laws we got from company-bribed senators and presidents. The US has a totally broken patent enforcement system, and it spreads the virus via WIPO, WTO, and so on.
Well, if companies are stupid enough to enter a joint venture WITHOUT making sure they get the rights to the patent too, then they are just plain incompetent. Also, a "secret patent" does not make any sense. Patents are published, patents MUST be published (this is the whole point of a patent). As for submarine patents, they are not kept in a secret drawer, they simply sit in the enormous USPTO archive, which is too big to be fully examined.
The original idea of patents is: publish your results, and we will grant you a temporary monopoly over it. The alternative? Companies keep their stuff secret.
So, assuming patents would not exist, the Africans would still be suffering from AIDS, the only difference being that the companies had kept their AIDS drugs a secret. (Good luck with reverse-engineering them.)
Patents are not a bad idea per se; its their enforcement and granting that are broken. AIDS drugs in Africa should pass as analogous to "fair use" clauses in copyright laws.
Of course, patenting does not make much sense for software..
EXT extensions are by all definitions standard and most of them do make it into ARB status after some time.
Yes, but as said, it is high time for a major GL cleanup. Today OpenGL is on-par with D3D9, and will be with D3D10, but it is a mess. 60-70% of all tutorials cover obsolete things like display lists, vertex arrays, or even drawing with glBegin/glEnd. Stuff like line stipples is included, now-useless extensions like ARB_imaging are still being dragged.. GL 3 is what everyone really wants - a clean cut. Only then one can really talk about GL being on par with D3D 10, since D3D 10 is massively-cleaned D3D, too.
actually DX has always had a batch problem, DX10 is like this to (only less so), but openGL doesn't really have that kind of extreme problem and openGL 3 will all but remove it (texture arrays will help too).
OpenGL *does* have a batch problem, but not as severe as D3D. The OpenGL ICD is smarter than the D3D HAL; the former has the ability to pre-batch and marshal the GL commands, the HAL can only pass commands through. However, this makes writing ICDs harder, and since OpenGL 3 is also aimed to be easier to support, maybe the ICD will be simplified too.
A technical preview is included in the DX9 SDK, and OpenGL3 won't be out before the Siggraph 07. New functionality might come via extensions, but a new, cleaned-up OpenGL is what everyone really wants and will get at Siggraph 07. In any way, D3D10 will be here earlier (hopefully GL won't be too late).
Besides, OpenGL already supports all and more features of the D3D part of DX10
Wrong. OpenGL only has an EXT extension for geometry shaders, but no superbuffers, texture arrays etc. so there is still much left.
(with better performance to boot)
Thats not the fault of Direct3D, its 100% a driver issue. nvidia cards are made for GL, hence the (slight!) performance advantage. On ATI cards, its totally different.
and vista doesn't even support DX10 yet since you need the DX10 graphics drivers that hasn't been released yet
You need new HARDWARE for this functionality, not just new drivers. Get a 8800.
Well, blowing holes into walls should be hard to achieve in the game, otherwise the map looks like swiss cheese after a while. Consequently, destructible environments are a nightmare for balancing.
Its not that you patented your specific technique. Its the fact that your patent is broad. So with it you block any chance for high-quality ragdolls in games unless they use your solution.
So, now, those who can, are not allowed to. Its that simple.
"Our technology for high-quality ragdolls is patented. This broad patent covers most spring/damper character simulation systems. If it falls, it has joints, it looks right, and it works right, it's probably covered by our patent."
There is no viable alternative. Theora is just too obscure to be useful. There are DirectShow codecs for Theora, but they don't work well, jump over frames, almost always mess up the A/V sync and so on. ffdshow theora is not enough, you need Ogg support too (since Ogg is the container, and theora is the video bitstream). Also, this involves installing codecs in 90% of all cases, since nobody has theora, as previously stated.
So. Alternatives? Dirac? Snow? Powerful and very advanced, but a) beta b) even less known.
So, for ensuring that everybody can watch a video, you have to go with WMV or Quicktime, or avi with mpeg4 video (divx, xvid). A very bad situation, indeed. But this is reality.
[i]If you aren't reasoning about the true nature of what you study, then you're not studying science any more. You've entered predictive religion.[/i]
Well, what is "truth"? The answer is: there is no "truth", at least not for science. Science deals with [i]models[/i]. *Religions* try to deal with truth. "There is a God waiting in heaven".... sold as absolute truth. How can you argue against an absolute truth? You see the dilemma here? Since science cannot claim to know the true nature of things, it deals with models applied to observable phenomena. Better models replace or extend current ones etc.
Once you try to nail down something as "THE truth", you get into trouble, or cause some for others.
Disclaimer: I am writing cross-platform games with backends for both Direct3D and OpenGL.
Now, its not hard with the right abstractions. C++ polymorphism suits very well for this. I encapsulated OpenGL and Direct3D in derivates of an abstract Rasterizer class. Works like a charm, and has the benefit of strongly separating the API from the rest of the code (which is a bad idea in 99,9% of all cases).
OpenGL caught up nowadays, VBOs are great, and FBOs allow rendertarget tricks beyond what D3D rendertargets can offer. Also, I consider GLSL to be superior to D3D HLSL. But I agree, when DX8 came out, OpenGL was is a sorry shape (no shaders, no vertexbuffers, only PBuffers for render-to-texture...)
I didn't know the java player, this is a nice project. But as for the DShow filters, these are the ones I meant. Forget them. Often they crash, and when they don't, the frames jump, or playback stops altogether. VLC is just a much better option. Also, "Updated: 2 May 2004" does not sound very good...
You forget that the geeks have nothing to say about this. This is a result of a patent and IP war. Video encoding is one of the most locked down areas of IT. So, no "gathering" of geeks will change anything.
One serious problem is that there are no working Theora VFW plugins (some exist, but they are unusable). You *have* to install VLC or mplayer. Add to this the fact that NO video editing software supports it, and you have a real mess (and don't start with mencoder -crypticoption1 -crypticoption2, I am talking about stuff like Premiere, or tools like VirtualDub. Oh, and there is no streaming server for Theora. This alone rules it out already.
Try some complex GL demos from Humus (www.humus.ca). Running them in Beryl freezes the machine using the nvidia 9631 drivers (no XGL/AIGLX, they are not needed because the driver handles this by itself).
Slashdot Mirror
Centralized repositories like the patent office.
You do know that mirrors exist?
Sure if you tell me why Ford would spend a billion to build a car factory even though GM could too.
Yet another pointless rant. Why does Ford spend the billion? Well, because they are the ones who will use it, they don't have to be afraid that GM will take over. Its their building, after all. It is, however, easy to simply rip off a novel technology if it is published.
Actually, IBM and Intell sued for billions over that and lost badly in the courts. So did the PC industry flop? no it boomed, no, in fact after that the IBM compatable PC took over the marketplace and there was a nuclear explosion in R&D, along with a nuclear explosion in business and commerce.
Yes, everyone gained from it, except IBM and Intel, who LOST. In the long term, they won too, but a scenario like this can easily break a company's neck. IBM was (and is) a big player, but imagine the PC had been invented by a small start-up company. They would have gone bankrupt.
The negros shouldn't need their "kind masters" permission to exercise freedom.
So your point is that scientists shall do their expensive and time-consuming R&D for free and not get a single cent paid for it, is that right? You do know that only government-funded research can afford this, right?
Oh, and cut the US crap already. I am not a US citizen.
Invention and creation happened 2000 years before patent, and got lost, forgotten, hidden. The antikythera knowledge was lost, and had to be re-invented, which wasn't guaranteed to happen. In fact, lost inventions may NEVER appear again. Imagine Goedel had never published his work, for example.
The free copy nature of the Internet is something COMPLETELY different. I am talking about inventions, not information. You are mixing patents with copyright.
Also, your problem seems to be that you don't consider how R&D is done in other departments. Equipment costs A LOT, lab costs are high, etc. You don't need that expensive equipment for writing software, you don't need a lab. Guess why FOSS didn't arise for hardware, or pharmaceuticals: BECAUSE R&D IS EXPENSIVE THERE. Getting rid of patents works only when R&D is comparatively cheap and a temporary monopoly is not plausible - which is very true for software, with the addition that major breakthroughs are very rare with software. However, pharma R&D is usually expensive. It *has* to pay off to do some research; how would it pay off if others can simply rip off my hard work? The only alternative would be to lock down the product so no one can reverse-engineer it; of course this also means that the company does not disclose the actual invention. If the company vanishes, the invention vanishes too. It already happened with stuff like Rotodyne.
So, now answer me this question: if R&D is (very) expensive, and everyone is free to rip the results, why should I do research in the first place? It just doesn't pay off!
Also, your comment about Africa shows that you simply didnt read what I said. Again: I AM ALL FOR FAIR USE CLAUSES. Africa should have gotten a permission in the first place.
Most of your points are valid only if you manually install the components. Distros like Ubuntu do take care of the 300MB libraries etc.
/usr/bin contains distro stuff and /usr/local/bin manually installed things. (local should be renamed, though.)
Some others:
- Today RPM and DEB are the de-facto standard package formats. There are also GUIs for installing these packages, also handling dependencies. (gdebi for DEBs, I dont know how the RPM one is named.)
- GStreamer is something like VFW, and is gaining popularity. (Ubuntu uses it by default.)
- The directory hierarchy is irrelevant for the home user. Again, take Ubuntu as example. Who cares about where the executables are? Click on the menu, and thats it. Also, its not hard to remember that
- The SuSE issue is definitely a design error. Note however that Windows has similar issues as well.
- Loading the libraries is slow, yes, partially because of missing symbol visibility (all symbols are loaded). KDE will make use of the gcc visibility feature to address this. As for Firefox, I don't know.
- *Advanced* administration is best done via command-line, because it has much greater flexibility than a GUI could ever have. Also, it scales better. Now, a list of installed codecs is available via gstreamer IIRC. Also, "bum" is a tool for administering the programs loaded at startup. It needs to be more known, though.
- I agree that there seems to be a bloat-mania in Linux land. Foobar2000 is *very* well coded. MPC, however, uses the VFW codecs, so the comparison with mplayer isnt really fair. vlc, mplayer, and xine seem to link everything, though (mplayer.exe is 9MB monster in Windows). Still, if I want to play *any* video, I use vlc, precisely because it has support for bazillion formats (even ones not supported in Foobar).
In sum, you are partially right, but as I said, most points are really a distro issue. Also keep in mind that most people do NOT install Windows by themselves; instead, they use what's on that Gateway PC they just bought. When Windows is borked again and infected with 15911 viruses after 6 months, Generic Geek is called and he sets up a fresh one. Or people call support, and get told to use the Recovery CD. With Linux it would be the same (it would be probably more robust, though).
So, the REAL reason why Linux isn't there yet is because it needs to be preinstalled on retail PCs.
You just don't get it, do you?
"massive restrictions on what people can copy" - Without patents, people could not copy anything BECAUSE THERE WOULD BE NOTHING TO COPY FROM. The science behind the drugs would be locked away deep in the company HQs. And no, people would NOT "invent them anyhow", precisely because the research is incremental. Company A makes small steps, patent them, B licenses A' patents, makes small steps... your alternative would result in stagnation because no one would be willing to disclose their results - everything would be secret. With patents, the results are PUBLIC.
Also, with patents, countries have the option to ignore them, like Brazil did. Keep in mind that some African countries are considering this as well. This is why I say that the enforcement is broken: there is no "fair use" clause for patents, save for usage by the military. THIS is what has to be changed. Africa is obviously in a state of emergency and should get free access to those patents.
As stated before, the AIDS problem is caused by a broken patent ENFORCEMENT. Without patents, the Africans *still* would not get any generics, except that the science behind them would be kept secret.
Blame the insane intellectual property laws we got from company-bribed senators and presidents. The US has a totally broken patent enforcement system, and it spreads the virus via WIPO, WTO, and so on.
Well, if companies are stupid enough to enter a joint venture WITHOUT making sure they get the rights to the patent too, then they are just plain incompetent. Also, a "secret patent" does not make any sense. Patents are published, patents MUST be published (this is the whole point of a patent). As for submarine patents, they are not kept in a secret drawer, they simply sit in the enormous USPTO archive, which is too big to be fully examined.
The original idea of patents is: publish your results, and we will grant you a temporary monopoly over it.
The alternative? Companies keep their stuff secret.
So, assuming patents would not exist, the Africans would still be suffering from AIDS, the only difference being that the companies had kept their AIDS drugs a secret. (Good luck with reverse-engineering them.)
Patents are not a bad idea per se; its their enforcement and granting that are broken. AIDS drugs in Africa should pass as analogous to "fair use" clauses in copyright laws.
Of course, patenting does not make much sense for software..
EXT extensions are by all definitions standard and most of them do make it into ARB status after some time.
Yes, but as said, it is high time for a major GL cleanup. Today OpenGL is on-par with D3D9, and will be with D3D10, but it is a mess. 60-70% of all tutorials cover obsolete things like display lists, vertex arrays, or even drawing with glBegin/glEnd. Stuff like line stipples is included, now-useless extensions like ARB_imaging are still being dragged.. GL 3 is what everyone really wants - a clean cut. Only then one can really talk about GL being on par with D3D 10, since D3D 10 is massively-cleaned D3D, too.
actually DX has always had a batch problem, DX10 is like this to (only less so), but openGL doesn't really have that kind of extreme problem and openGL 3 will all but remove it (texture arrays will help too).
OpenGL *does* have a batch problem, but not as severe as D3D. The OpenGL ICD is smarter than the D3D HAL; the former has the ability to pre-batch and marshal the GL commands, the HAL can only pass commands through. However, this makes writing ICDs harder, and since OpenGL 3 is also aimed to be easier to support, maybe the ICD will be simplified too.
A technical preview is included in the DX9 SDK, and OpenGL3 won't be out before the Siggraph 07. New functionality might come via extensions, but a new, cleaned-up OpenGL is what everyone really wants and will get at Siggraph 07. In any way, D3D10 will be here earlier (hopefully GL won't be too late).
Besides, OpenGL already supports all and more features of the D3D part of DX10
Wrong. OpenGL only has an EXT extension for geometry shaders, but no superbuffers, texture arrays etc. so there is still much left.
(with better performance to boot)
Thats not the fault of Direct3D, its 100% a driver issue. nvidia cards are made for GL, hence the (slight!) performance advantage. On ATI cards, its totally different.
and vista doesn't even support DX10 yet since you need the DX10 graphics drivers that hasn't been released yet
You need new HARDWARE for this functionality, not just new drivers. Get a 8800.
Well, blowing holes into walls should be hard to achieve in the game, otherwise the map looks like swiss cheese after a while. Consequently, destructible environments are a nightmare for balancing.
Its not that you patented your specific technique. Its the fact that your patent is broad. So with it you block any chance for high-quality ragdolls in games unless they use your solution.
So, now, those who can, are not allowed to. Its that simple.
From your site:
"Our technology for high-quality ragdolls is patented. This broad patent covers most spring/damper character simulation systems. If it falls, it has joints, it looks right, and it works right, it's probably covered by our patent."
Thank you for stifling innovation yet again.
There is no viable alternative. Theora is just too obscure to be useful. There are DirectShow codecs for Theora, but they don't work well, jump over frames, almost always mess up the A/V sync and so on. ffdshow theora is not enough, you need Ogg support too (since Ogg is the container, and theora is the video bitstream). Also, this involves installing codecs in 90% of all cases, since nobody has theora, as previously stated.
So. Alternatives? Dirac? Snow? Powerful and very advanced, but a) beta b) even less known.
So, for ensuring that everybody can watch a video, you have to go with WMV or Quicktime, or avi with mpeg4 video (divx, xvid). A very bad situation, indeed. But this is reality.
Argh ... damn bbcode habits.
[i]If you aren't reasoning about the true nature of what you study, then you're not studying science any more. You've entered predictive religion.[/i]
Well, what is "truth"? The answer is: there is no "truth", at least not for science. Science deals with [i]models[/i]. *Religions* try to deal with truth. "There is a God waiting in heaven".... sold as absolute truth. How can you argue against an absolute truth? You see the dilemma here? Since science cannot claim to know the true nature of things, it deals with models applied to observable phenomena. Better models replace or extend current ones etc.
Once you try to nail down something as "THE truth", you get into trouble, or cause some for others.
Disclaimer: I am writing cross-platform games with backends for both Direct3D and OpenGL.
Now, its not hard with the right abstractions. C++ polymorphism suits very well for this. I encapsulated OpenGL and Direct3D in derivates of an abstract Rasterizer class. Works like a charm, and has the benefit of strongly separating the API from the rest of the code (which is a bad idea in 99,9% of all cases).
OpenGL caught up nowadays, VBOs are great, and FBOs allow rendertarget tricks beyond what D3D rendertargets can offer. Also, I consider GLSL to be superior to D3D HLSL. But I agree, when DX8 came out, OpenGL was is a sorry shape (no shaders, no vertexbuffers, only PBuffers for render-to-texture...)
Wrong.
Superbuffers? Texture arrays? Geometry shaders? Constant buffers? FFP gone? Improved instancing? Predicated geometry?
Mind you, OpenGL 3 will have these, too.
This has been disproved a long time ago. NVidia and ATI install their own OpenGL ICDs just fine.
I didn't know the java player, this is a nice project.
But as for the DShow filters, these are the ones I meant. Forget them. Often they crash, and when they don't, the frames jump, or playback stops altogether. VLC is just a much better option. Also, "Updated: 2 May 2004" does not sound very good...
Besides, Premiere still doesn't support theora...
You forget that the geeks have nothing to say about this. This is a result of a patent and IP war. Video encoding is one of the most locked down areas of IT. So, no "gathering" of geeks will change anything.
What alternatives do we have? Dirac? Snow?
Can these be inserted into the ogg bitstream without too much pain?
One serious problem is that there are no working Theora VFW plugins (some exist, but they are unusable). You *have* to install VLC or mplayer. Add to this the fact that NO video editing software supports it, and you have a real mess (and don't start with mencoder -crypticoption1 -crypticoption2, I am talking about stuff like Premiere, or tools like VirtualDub. Oh, and there is no streaming server for Theora. This alone rules it out already.
Try some complex GL demos from Humus (www.humus.ca). Running them in Beryl freezes the machine using the nvidia 9631 drivers (no XGL/AIGLX, they are not needed because the driver handles this by itself).