Inputs without Outputs: What to do with them?

Fishkin’s taxonomy elevates one’s awareness of a question that’s rarely asked, but worth asking: what exactly comprises a TUI, and how can we systematically compare them to each other? That is, we know a TUI when we see one, and we can easily tell them apart, but Fishkin’s text shows us one way of doing so–and I confess that, while reading, I found myself jotting down new ideas provoked by the rubric. So I did find it productive and helpful, in short.

That said, I wonder about a corner not explicitly covered in the text. Fishkin remarks in the section “Calm Computing” on TUIs stripped of their human inputs: a flower’s petals respond to ambient temperature; or a self-shortening string that responds to bus wait times. These examples are “tangible,” I suppose, insofar as they have components that could conceivably be touched, but they don’t demand user input. Can we imagine, however, a device that would have user input and not produce an output? Can we properly categorize such a device as a TUI?

Take, for example, a button that says “press,” but which has no effect. This button would seem to fall into an unclear categorization in terms of embodiment: would it comprise “full” embodiment insofar as the pushing of the button is itself the user experience? Or would it engage the user as “nearby” or “distant” embodiment (depending on the user’s expectations) that never comes to fruition? What about metaphor: is the metaphorical relationship of the experience “none” because the pushing of the button lacks any kind of relationship with the outcome (where pushing buttons classically results in effects), or would the interaction be better described as “full”? Clearly the example is problematic, and we might think of any number of other examples: a light switch that a user flips while simultaneously knowing it connects to nothing; a television screen that a child touches as if it were touch-sensitive; a synthesizer that one plays while it is off. How would Fishkin’s taxonomy account for such examples—and if it cannot, what categories might we add to help it succeed in this way?


Leave a Reply