A rigorous and general theory, as well as an associated efficient computational methodology, was developed that addresses the question of when and what linear control is adequate for a nonlinear process. A number of computer simulation examples illustrate the power of the proposed approach. Several potential future developments are outlined.