This paper gives the author's views on the kind of computer-based modeling environment needed to properly support management science/operations research work, and on the design challenges that need to be met in order to bring such modeling environments into being. It is a written version of the main ideas of two addresses: a plenary at