Presented in this paper are some preliminary ideas on the role of actuator saturation in linear systems that are under study. Real actuators have limited bandwidth and they saturate upon receipt of overly large inputs. Such actuator saturation can have catastrophic consequences especially in feedback control of open-loop unstable plants. It is well known that even when used with stable plants, actuator saturation can lead to undesirable limit cycle behavior. Actuator saturation can also easily occur when integrations are used within a control loop. Integrator anti-windup schemes have been proposed for overcoming that situation. One obvious but almost unresearched idea for eliminating wind up would be to get rid of the need to use integral action inside the feedback loop. So the crucial question is whether or not it is possible to get good steady state accuracy without using integral control? One idea could be to delegate the tracking responsibility to a carefully designed pre-filter.