ATI doesn't have the same number of Direct X 10 cards as nVidia. As per the Valve Hardware Survey, ATI has about 13 000 Radeon 2600 HD cards installed. nVidia, on the other hand, has 137 000 8800's installed alone. Thats not counting any other 8000 series cards. In other words, of course nVidia is looking worse, there's more opportunity for something to go wrong. I don't blame anyone for the graphics card crashes really; the changes from Direct X 9 to Direct X 10 are quite drastic. I just wish Vista's release hadn't been rushed so that everyone could get their shit together.
America's Army is kinda different, because the training is administered in such a way that they basically take an actual training course and present that to a person. In other words, the way Americas Army presents it is no different from an actual course in except for the depth. America's Army uses real pictures with real people (of course faked I think) and a course that, while simplified, teaches basic first aid (Controlling bleeding, Clearing and maintaining airways) in a fairly effective manner
The real issue is that skills like this aren't taught any way other than through specialized activities. You'd think that keeping children safe would include teaching them what to do after they find something unsafe, but no.
This situation is already encountered with rasterized methods and has been solved. The shadows have a limit, beyond which they will not be cast. Another part of the solution is that shadows are only cast using the highest resolution mesh.
AS much as I would like to see raytracing, There's no way that we will see it take over rasterized methods for quite some time. For one thing, elements of what you have said have already come. Stencil shadows using John Carmacks depth-fail method already provide pixel perfect shadows. While glass currently cannot perform realtime cubemap rendering, in the future I can see glass being able to render true reflections.
All this is additionally limited because most of the Realtime raytracing techniques I have seen are not able (currently) to use shader programs or anything similar. Vertex shaders, which modify aspect of vertices, have obvious uses in all areas of 3d rendering. The biggest issue I see is the lack of fragment shaders (pixel shaders in Direct X). These allow complex shading, such as real-time radiosity, real-time ambient occlusion, Relief mapping, Post processing, etc.
Despite all this, Raytracing will eventually overtake rasterization due to raytracing scaling to cores much better that rasterization. It is hard to predict how soon we'll see raytracing, and to be honest I don't see it happening too soon (unless nVidia and AMD/ATI release raytracing optimized graphics cards)
Slashdot Mirror
ATI doesn't have the same number of Direct X 10 cards as nVidia. As per the Valve Hardware Survey, ATI has about 13 000 Radeon 2600 HD cards installed. nVidia, on the other hand, has 137 000 8800's installed alone. Thats not counting any other 8000 series cards. In other words, of course nVidia is looking worse, there's more opportunity for something to go wrong. I don't blame anyone for the graphics card crashes really; the changes from Direct X 9 to Direct X 10 are quite drastic. I just wish Vista's release hadn't been rushed so that everyone could get their shit together.
America's Army is kinda different, because the training is administered in such a way that they basically take an actual training course and present that to a person. In other words, the way Americas Army presents it is no different from an actual course in except for the depth. America's Army uses real pictures with real people (of course faked I think) and a course that, while simplified, teaches basic first aid (Controlling bleeding, Clearing and maintaining airways) in a fairly effective manner
The real issue is that skills like this aren't taught any way other than through specialized activities. You'd think that keeping children safe would include teaching them what to do after they find something unsafe, but no.
This situation is already encountered with rasterized methods and has been solved. The shadows have a limit, beyond which they will not be cast. Another part of the solution is that shadows are only cast using the highest resolution mesh.
AS much as I would like to see raytracing, There's no way that we will see it take over rasterized methods for quite some time. For one thing, elements of what you have said have already come. Stencil shadows using John Carmacks depth-fail method already provide pixel perfect shadows. While glass currently cannot perform realtime cubemap rendering, in the future I can see glass being able to render true reflections.
All this is additionally limited because most of the Realtime raytracing techniques I have seen are not able (currently) to use shader programs or anything similar. Vertex shaders, which modify aspect of vertices, have obvious uses in all areas of 3d rendering. The biggest issue I see is the lack of fragment shaders (pixel shaders in Direct X). These allow complex shading, such as real-time radiosity, real-time ambient occlusion, Relief mapping, Post processing, etc.
Despite all this, Raytracing will eventually overtake rasterization due to raytracing scaling to cores much better that rasterization. It is hard to predict how soon we'll see raytracing, and to be honest I don't see it happening too soon (unless nVidia and AMD/ATI release raytracing optimized graphics cards)