Through the Computing Research Policy weblog, I came across this attempt to define potentially high-impact areas of computer science. As of now, the big four are: internet, search, cluster computing and computational science.
I am especially pleased to see the last entry and the implication that computer science is more than just the study of computing machinery or computational processes in the abstract. As stated in that blog, the focus still seems to be on a conventional view of using large computational simulations to aid science (as is common in, say, material science or meteorology) but I am choosing to interpret this theme in a broader sense as what some people in Edinburgh call ‘computational thinking‘. The idea is to focus on the doing of science (biology, economics, etc.) with a CS/engineering hat on – in a way that is different from other traditional ways of doing that science. An example (deliberately picked from outside the oft-repeated ones) I have been thinking about lately – can we use autonomous learning agents in artificial markets to advance our understanding of concepts in economic theory such as bounded rationality or the role of asymmetric information? A lot of computational work in economics is either done in the old-style pen and paper format or in the behavioural and empirical format. How about a third format – writing computer programs in virtual worlds and doing this well enough to have tangible impact on the field?