The minimal thing I think would be necesary to make OpenGL an better scene graph is to have an display list reference with a single transform matrix. Hierarchal information (relativity of objects) may not be necesary or may be an add on kind of utility (glut maybe). The import information for a renderer to have is the location of objects, especially when it comes to real time shadow and raytracing effects (which if done today would require accumulation of meshes every frame and even then object clipping is not always done by OpenGL implementations).
I agree about animation for the time being because I dont think we have hit the highest parts of this yet, but any one who is familiar with Softimage|3D or Maya knows that animation can do some really nice things. I think when boned animations become more useful that we will start to see these things in games more often, current games are still using moprhs for animations.
I shouldnt discount programmibility out right, I think some very interesting things could be developed with it (although I'm not sure how extensive current implementations are), it would be nice to start seeing procedural texture, objects, and fog effects. But procedural is a good option when you are low on memory, but memory is cheap these days, I think it would be nice if an API allowed for real time shadows in at least a minimal scene graph.
Another interesting thing that scene graphs allow for is optimizations. Programmers use several techniques these days, like BSPs Quad/OctTrees and Portals, in order to optimize rendering and collision detection. While they could still be used to optimize collision detection, it could be allowed that a video system can optimize a scene to the way it prefers to render it.
The problem is copyrights cover more then just giving credit for work, and so I think the reason people don't give credit is the fear of liability that comes with it. (I think this case may be just that)
For example if I use parts of someone elses source code and give them credit. The author I give credit to for those parts may in turn make demands about the copyright license I use. Worst case scenero they may demand I use a particular license (like BSD) for all the source code or not use their source code at all. Another one is they may require me to seperate the pieces of their code and put it under their license, there by preventing any benefits/performance or other of integration of code (imagine if everyone made such demands and each had their own license, we would see the open source world as seperated as the world is today).
So anyway my point is, copyrights cause a lot of trouble, they reduce the chance someone is going to get credit for something (that can be obfiscated or altered so easily) due to fear of the liability that comes with it, there by aiding and influencing in the destruction of history of ideas and source. Same with reverse engineering in some cases, where reverse engineering is prevented, in which case someone who gets something from reverse engineering will not likely give credit to it, because of the chance they could get in trouble.
If copyright was merely forcing people to give credit for work they have done, there would not be a problem.
All work originates from nature, things are not created in a vacuum, they are created from other things from nature and from other peoples works (which comes from nature or other peoples work). Its nice to have a history of where something came from and of the interest of the individual to have claim they did something, but we should not forget that everything comes from something else and our minds simply filter and alter things from their origin in order to create something new and/or useful (like clay). These things should not be forgotten for the sake of the individual, but should be remembered for the sake of every individual.
In my opinion, they seem to focus a lot on programmability, but I think the future is in the SceneGraphs not so much the programmability. I think they are thinking to low level, when there needs to be many changes to the high level part of it. With high level programming for example shadows dont have to be customized hacks, and its even possible for the user to customize to their system what kind of effects to use. Scene graphs also give the video card more information about a scene necesary for many realistic effects.
While OpenGL does have partial scene graph support I think a lot more could be done, a lot of redundant tasks that programmers have to do in creating a more complete scene graph.
Another area is animation. Supporting 2d graphics is one thing, supporting 2d animation/video is another aspect that I have heard considered, and that can also be extended to 3d animation in conjunction with better scene graph support.
It just seems to me that, to many of the vendors are looking for low level programmers. Most game developers these days are starting to reduce redundancy, rather then reinventing the scenegraph, they are starting to use for example the Unreal or Quake Engine, which is really a larger package of a scene graph and a whole lot more. And I think the reason they look at it this way, is that scene graphs would take a lot of time and investment on the part of the IHVs drivers then on the game and application developers.
I agree with you in most part but you sort of contradict yourself here
It's not illegal to sell someone else's parts if you acquired them legally, but it is illegal to claim them as your own make without permission, regardless of how they were acquired.
Like aquiring via reverse engineering? I know that its not illegal to use reverse engineering to make something compatible. But I have to question of whether Søren actually owns the structures or not, I dont believe he does own them, but then again I only believe in 'Intellectual property' as being a property of the media (and the media the property of whom ever posseses it) not the creator. I do think Søren should recieve some credit for his part (a bibliography of sorts), but only on the level that not only credit but also a path from which the knowledge came about (history) so that those doing research may learn from the source, I dont like copyrights, I dont think its governments place to force people how to think. I think copyrights are the exact reason why no one gives anyone else credit, for example if I dont like the BSD license, I am more likely not to give credit to the author because that could make me liable to release the code I created under BSD license, plus there are issues of having pieces of code being one license and other pieces of code in other licenses. It sounds like a horrible mess to me. If we didnt have copyright licenses or restrictions on reverse engineering, people would be more willing to document their sources because there would be no liability in doing so (like forcing confessions). I dont mean to sound arguementative but this is how I see things.
Slashdot Mirror
The minimal thing I think would be necesary to make OpenGL an better scene graph is to have an display list reference with a single transform matrix. Hierarchal information (relativity of objects) may not be necesary or may be an add on kind of utility (glut maybe). The import information for a renderer to have is the location of objects, especially when it comes to real time shadow and raytracing effects (which if done today would require accumulation of meshes every frame and even then object clipping is not always done by OpenGL implementations).
I agree about animation for the time being because I dont think we have hit the highest parts of this yet, but any one who is familiar with Softimage|3D or Maya knows that animation can do some really nice things. I think when boned animations become more useful that we will start to see these things in games more often, current games are still using moprhs for animations.
I shouldnt discount programmibility out right, I think some very interesting things could be developed with it (although I'm not sure how extensive current implementations are), it would be nice to start seeing procedural texture, objects, and fog effects. But procedural is a good option when you are low on memory, but memory is cheap these days, I think it would be nice if an API allowed for real time shadows in at least a minimal scene graph.
Another interesting thing that scene graphs allow for is optimizations. Programmers use several techniques these days, like BSPs Quad/OctTrees and Portals, in order to optimize rendering and collision detection. While they could still be used to optimize collision detection, it could be allowed that a video system can optimize a scene to the way it prefers to render it.
The problem is copyrights cover more then just giving credit for work, and so I think the reason people don't give credit is the fear of liability that comes with it. (I think this case may be just that)
For example if I use parts of someone elses source code and give them credit. The author I give credit to for those parts may in turn make demands about the copyright license I use. Worst case scenero they may demand I use a particular license (like BSD) for all the source code or not use their source code at all. Another one is they may require me to seperate the pieces of their code and put it under their license, there by preventing any benefits/performance or other of integration of code (imagine if everyone made such demands and each had their own license, we would see the open source world as seperated as the world is today).
So anyway my point is, copyrights cause a lot of trouble, they reduce the chance someone is going to get credit for something (that can be obfiscated or altered so easily) due to fear of the liability that comes with it, there by aiding and influencing in the destruction of history of ideas and source. Same with reverse engineering in some cases, where reverse engineering is prevented, in which case someone who gets something from reverse engineering will not likely give credit to it, because of the chance they could get in trouble.
If copyright was merely forcing people to give credit for work they have done, there would not be a problem.
All work originates from nature, things are not created in a vacuum, they are created from other things from nature and from other peoples works (which comes from nature or other peoples work). Its nice to have a history of where something came from and of the interest of the individual to have claim they did something, but we should not forget that everything comes from something else and our minds simply filter and alter things from their origin in order to create something new and/or useful (like clay). These things should not be forgotten for the sake of the individual, but should be remembered for the sake of every individual.
In my opinion, they seem to focus a lot on programmability, but I think the future is in the SceneGraphs not so much the programmability. I think they are thinking to low level, when there needs to be many changes to the high level part of it. With high level programming for example shadows dont have to be customized hacks, and its even possible for the user to customize to their system what kind of effects to use. Scene graphs also give the video card more information about a scene necesary for many realistic effects.
While OpenGL does have partial scene graph support I think a lot more could be done, a lot of redundant tasks that programmers have to do in creating a more complete scene graph.
Another area is animation. Supporting 2d graphics is one thing, supporting 2d animation/video is another aspect that I have heard considered, and that can also be extended to 3d animation in conjunction with better scene graph support.
It just seems to me that, to many of the vendors are looking for low level programmers. Most game developers these days are starting to reduce redundancy, rather then reinventing the scenegraph, they are starting to use for example the Unreal or Quake Engine, which is really a larger package of a scene graph and a whole lot more. And I think the reason they look at it this way, is that scene graphs would take a lot of time and investment on the part of the IHVs drivers then on the game and application developers.
I agree with you in most part but you sort of contradict yourself here
It's not illegal to sell someone else's parts if you acquired them legally, but it is illegal to claim them as your own make without permission, regardless of how they were acquired.
Like aquiring via reverse engineering? I know that its not illegal to use reverse engineering to make something compatible. But I have to question of whether Søren actually owns the structures or not, I dont believe he does own them, but then again I only believe in 'Intellectual property' as being a property of the media (and the media the property of whom ever posseses it) not the creator. I do think Søren should recieve some credit for his part (a bibliography of sorts), but only on the level that not only credit but also a path from which the knowledge came about (history) so that those doing research may learn from the source, I dont like copyrights, I dont think its governments place to force people how to think. I think copyrights are the exact reason why no one gives anyone else credit, for example if I dont like the BSD license, I am more likely not to give credit to the author because that could make me liable to release the code I created under BSD license, plus there are issues of having pieces of code being one license and other pieces of code in other licenses. It sounds like a horrible mess to me. If we didnt have copyright licenses or restrictions on reverse engineering, people would be more willing to document their sources because there would be no liability in doing so (like forcing confessions). I dont mean to sound arguementative but this is how I see things.