I think C2D is only faster because the transfer rate from AGP to CPU (NB bottleneck) is much higher. For instance, on my XP platform, I see no scaling from increasing the CPU frequency, but huge scaling when I increase the FSB frequency.
It was removed because it was highly inefficient: too much saved records for the purpose of the feature.
I think RB already stated that a new version is planned, although ... let us fix all other bugs before that.
The idea of the oc challenge is to have one stage a month, not two or subrankings or anything like that. If we start doing this, we end up having to split up the competition in a number of rankings equal the the amount of hardware categories we have ... which makes it too chaotic an uninteresting.
The reason why the chosen platform is so old is to allow interested people to go to ebay and find a card to participate ;-)
That would mean that we litterally equalize "effort/skill" and "able to buy hardware" in terms of points and reward. I'm sure that's not how any ranking should be.
A steep curve is needed to differentiate those who actually break the world record by going all LN2 from those who want to come close only using air vga's.
Absolute WR = absolute WR.