When reading this definition of interactivity, I can easily see a modification of it that clarifies it some in my mind: “two actors alternatively reacting to each other”. This clearly requires a definition to react, but that is most obviously the same as what it replaced in the original definition, input, process, output. His list of things that are not interactive are simply things that are not reacting to the person he is describing as the other actor. The scale of interactivity then boils down to the complexity of the reactions, with a twist. The twist is that one of the actors n every situation he mentions is a person, and when the other is a machine, the main target of this analysis, it is only the complexity of the reactions from the perspective of the user that matters. This is where is interactive and interface designer seem to differ. The interface designer hides the “unnecessary” parts of the reaction, while the interactive one reveals as much as possible to emphasize the complexity. By the definitions that I am using the reasons the interactive approach are obvious: making the machines reaction seem more complex make it seem more interactive, and the users reaction to this additional information actually do increase the interactivity, at the cost of efficiency of the primary task. That is unless the primary task is to get the user to think. With our definitions this could be rephrased as increase the complexity of reaction from the human user, which is a clear consequence of increasing the reactivity of the machine, an makes clear, at least to me, what makes an experience interactive and why that might be valuable.