The intent of this talk is to impart to controls researchers in academia and industry an appreciation of application-oriented considerations for nonlinear control. I will pay specific attention to process industry applications, although other application domains will also be mentioned. Difficulties of identification, multivariable complications, some shortcomings of "intelligent control", and the lack of provable guarantees are discussed in this context. Yet if such challenges can successfully be met, exciting possibilities lie ahead. Multi-modeling architectures and randomized algorithms, among other topics, are key research topics to this end. Concepts for large-scale optimization, intelligent data analysis, and complex adaptive systems are outlined as potential breakthrough applications. I conclude by offering some reflections on defining "nonlinear control", in the hope of broadening the connotation of the term within the controls community.