Futuremark was evidently trying to avoid another fiasco by allowing only specific drivers that they have tested and approved. However, their approval and recommendation of the NVIDIA Rev. 52.16 driver with its 3DMark03-specific optimization seems to run counter to that objective.
Granted, they tested NVIDIA's optimization and found that it only affected the Pixel Shader 2.0 test and had no effect on the overall 3DMark score. Still, that should not have been the ONLY approved driver on the list.
If there has to be a single approved driver, it should have been NVIDIA's latest "optimization-free" driver. That would allow everyone to produce valid and completely comparable results.
Even if they felt that the Rev. 52.16 driver was still a valid driver, why not provide a list of several approved drivers including at least one "optimization-free" driver? After all, they already list three different approved drivers for ATI cards.
Del med dine venner