Lambda the Ultimate

Hi everyone,

This is my first post here, so I hope I will not mess it up..
I am looking for a paper which I might probably have imagined that I have seen mentioned in a thread here once. It was about the optimal algorithm, a generic one in some sense. IIRC, it is circa 1999-2001.
Actually the general idea I am looking for is that a program which is optimized for time cannot be too bad in terms of memory and vice versa. I think this was mentioned on that idea but I would be happy to get any other references. I am interested in what this might mean for machine learning (I am doing my PhD in this area).
Can you please inform me of this paper or any relevant ones, if possible?

Thanks,

Muhammad Alkarouri