This paper describes the initial work in a program of research that aims to create diagrammatic tools for the modeling and discovery of an extremely challenging domain: Complex multi-dynamic (CMD) processes. Diagrams play a central role in the approach because they provide a range of representational, cognitive and computational properties that may yield potential solutions to the challenges posed by CMD processes. Motivation for this work comes from previous research on the use of diagrams in computational scientific discovery [2], [3] and the potential of diagrams for conducting knowledge acquisition [4].