I don't mean to be flippant. I know to many of you the difference between one mip-map level of detail and another is the difference between life and death. But until graphics APIs define precisely what outputs should be the result of each call, we'll have IQ differences, and they will continue becoming increasingly hard to pin down. We will have to live with some ambiguity, as difficult as that is for a community of folks addicted to extensive empirical testing and sweet, sweet graphical results. Even though I spend hours and hours of my time testing such things, I am comfortable living with some degree of ambiguity, when I must. I know not everyone is.
That said, I think we should hold graphics companies responsible for as much as possible, and I believe we've seen some clear violations of public trust and, ahem, invidious behavior. We will continue to point such things out.
But I may not get that IQ comparison done as promised, in large part because I found so little that's really interesting to report. As always, we'll roll the useful results of my efforts into future articles around hereand soon.