So, what we knew:

>UE3 has not any AA method implemented, and Batman:AA is coded over UE3.

>NVIDIA has helped the developers of B:AA to implement a custom AA filter.

>That custom AA filter developed in colaboration with Eidos is nothing exclusive to NVIDIA, since it's been proven to work with other D3D cards.

>Since NVIDIA has helped to implement (or completely implemented) that AA filter, they feel with the right to not allow people with hw of other IHVs to run their code, and "encourage" other vendors to implement their own code if they want that feature running on their hw, even when that code is perfectly compatible with any standard hw.

That bring us again to the main point of the discussion until now: is that right? Where those practices lead us, the consumers?

It would be the same thing if DIRT2 DX11 features (or part of them) don't work on NVIDIA DX11 hw when they release any, since AMD has supported the developement of that features.

It would be the same thing if OpenCL acceleration of Havok only works on AMD and Intel hw for the same reason.

The same for OpenCL acceleration in Bullet Physics.

Great for all consumers. There was a time where software coded over standard interfaces could be run on any hardware compliant with those standards. That was the whole point of those standards. There was a time... thanks, NVIDIA.