The paper discusses alternative ways to think about the modeling endeavor; the importance of including qualitative factors (i.e., "soft factors") despite critics who think doing so reduces rigor; the fundamental necessity of worrying seriously about uncertainty from the outset, including the kinds of uncertainty present in complex adaptive systems; and about implications for design of models. The paper's admonitions would be straightforward except that they fly in the face of common organizational practice, which is to avoid soft factors, ignore uncertainty by obsessing on standard cases, and use models ill-designed for serious uncertainty analysis.