Historically at UW-Madison, researchers have been fairly self-reliant for their IT service needs.  As infrastructure becomes more complicated and interdependent, technical support becomes more costly, and various components of the cyberinfrastructure can be more easily leveraged; that model is apt to change a bit. At a recent meeting, Dr. Terry Millar from our graduate school made an interesting observation.  He said, "to understand how IT service can best be applied to help our researchers, we need to understand that there are really three components: 1) infrastructure, 2) the science of infrastructure, and 3) the science of science.  This observation has helped me think about strategies we might embrace as we move forward and should help us all frame how best to apply IT service for UW-Madison research.