I can't believe so many people think this is excusable. The option to force trilinear mipmaps used to be in the drivers and there is still hardware support for it. However, they took the option to force them away.
As for making old games run perfectly or fairly close to it, there is also no excuse for that either because of shaders and stuff like CUDA and OpenCL... anything can be emulated with 100% (or damn near 100%) accuracy and decent speed if double floating point precision (or in some cases full single fixed point shader precision will be sufficient) is used in the shaders... in addition, they can use spoof video card I.D.s. They could even create a virtual device based upon older hardware or they could even make a virtual device that was an improvement of old hardware. I think a Geforce FX with added features like rotated grid AA and a fully orthogonal rendering pipeline emulated in double fp precision would be an excellent idea. It could even use a Geforce FX spoof I.D. OTOH, the GTX680 doesn't have very good FP64 shader performance, but maybe when the GTX780 comes out they won't cripple FP64 shader performance and instead we can have 512 GFLOPs double FP precision performance... and that should be enough to play the original Splinter Cell the way it was meant to be played (i.e., with shadow buffers and the better lighting quality) and with 4x RGMS (or RGSS) to boot (no pun intended).
As for making old games run perfectly or fairly close to it, there is also no excuse for that either because of shaders and stuff like CUDA and OpenCL... anything can be emulated with 100% (or damn near 100%) accuracy and decent speed if double floating point precision (or in some cases full single fixed point shader precision will be sufficient) is used in the shaders... in addition, they can use spoof video card I.D.s. They could even create a virtual device based upon older hardware or they could even make a virtual device that was an improvement of old hardware. I think a Geforce FX with added features like rotated grid AA and a fully orthogonal rendering pipeline emulated in double fp precision would be an excellent idea. It could even use a Geforce FX spoof I.D. OTOH, the GTX680 doesn't have very good FP64 shader performance, but maybe when the GTX780 comes out they won't cripple FP64 shader performance and instead we can have 512 GFLOPs double FP precision performance... and that should be enough to play the original Splinter Cell the way it was meant to be played (i.e., with shadow buffers and the better lighting quality) and with 4x RGMS (or RGSS) to boot (no pun intended).
Last edited: