Quoted By:
Very roughly speaking, computer science tends to deal with declarative problems, that is, how can I find a solution to x problem/class of problems, what steps do I need to take, how will I know the answer is good enough, etc. Wheras mathematics is more imperative, that is stating a definition and exploring the implications that necessarily follow from that definition and the agreed on rules of inference.
For example, consider the problem of finding a square root. Mathematics is saying that Note that we give no instructions on how to find or compute the value of x, only stated what it is. Computer Science is the discipline of exploring "how" we can compute the square root of a number, how to do it effectively (is our solution correct?) and efficiently (how much memory does it take? How long does it take to compute our value?), and also dealing with representing and dealing with infinite, continuous objects with discrete elements and methods (a computer is ultimately relying on 1's and 0's, after all)
And then Physics is trying to match mathematical models to real world observations, and using mathematical ideas to model and predict real world phenomena. In pure mathematics we often work with the ideal, perfect world, but in reality things are rougher and our measurements have limitations with regards to precision and accuracy.
I want to emphasize again that this is a very rough and broad description and in reality the disciplines have lots of crossover. Depending on the sub discipline and topic, there is lots of overlap between the 3. For example computer science often uses methods from calculus (a topic that deals with continuous structures), or discrete analogues of the same ideas that are more or less equivalent.
