Historically, activity-centered ergonomics approach has developed the method of Ergonomics Work Analysis, the aim being to define the problems encountered by the workers in an existing work setting. But the active participation of ergonomists in the design of large-scale industrial or informatics projects at the beginning of the 1980s will lead to question simulation as a unavoidable method due to the need to producing knowledge about situations that do not yet exist. We will argue that these methods have moved from a “figurative” to an “operative” approach, which can be summarized in six points: 1. Simulation is a ubiquitous method during design. However due to its concern (the living, cultural and social work activity), ergonomics is confronted with epistemological questions different from those of the engineers. This specificity leads to relativizing, and even abandoning any predictive ambition. To simulate is to stage the coupling between worker(s) and object(s) to design, in order to manipulate and to understand it. 2. Understanding simulation as a method explicitly apprehended and thought in relation to a future situation is not sufficient. To simulate is to define a specific and finalized work setting, which must be apprehended and defined in the specificity of its cognitive, cultural and social dimensions. 3. Questions such as "ecological validity" do not disappear, but this validity must be situated in the context of the design process. For example, methods developed with theoretical modesty (eg mock-up associates the users) may appear preferable to more scientifically or technically founded methods (eg cognitive simulation, full-scale simulation, …), because they are better suited to the temporal dynamic of the design process. 4. Even if simulation is not explicitly serving as a learning goal, the use of a model appears undoubtedly associated to any questioning on learning and development. Simulation brings on learning in two ways: (i) through the building of a model, and (ii) in using the model. 5. The use of a model involves collective and multi-voices activities. For ergonomists, the challenge is that simulation methods contribute to a mutual learning process between users and designers. Mutual learning can be grasped as dialogical processes, where the result of the activity of one actor (designer or user), implemented though a model, will be validate, refute, or set in motion, based on activity performed by an other actor involved in the process. From that point of view, simulation is intrinsically a participative method. 6. Simulation methods should provide resources for operators to build a model of their future work setting. The aim for the workers is to focus on their own activity, to conduct an investigation in order to identify what is causing them problems and troubles, and to experience possible resolution to depict.