This paper addresses the steady-state optimization of continuous processes in the presence of uncertainty in the form of unknown or time-varying model parameters, structural plant-model mismatch, and disturbances. To address these issues, we assume that certain measurements are available in real time, and the question is how the plant inputs can be updated based on these measurements to drive the plant to optimality. This is the field of static real-time optimization (RTO). This paper shows that there are two distinguishing features that have a significant impact on the formulation and performance of RTO schemes. The first of these features is the type of measurements that is used, either generic or specific. The former corresponds to routinely available plant outputs, while the latter includes optimization-specific information such as the cost and constraint values with appropriate input excitation. The second of these features is the manner in which optimality is implemented, that is, via either repeated optimization or feedback control. It is argued that, if specific information is available, plant optimality can be achieved irrespective of the quality of the model. Also, the implementation method must consider the speed at which the information can be collected and the computational time required for numerical optimization. The importance of these implementation aspects is illustrated in a simulation of the optimization of the Willam-Otto reactor.