Ju, Lee, and Klemmer (CSCW 2008) discuss implicit and explicit interaction using the example of a vertical interactive surface in the Range application. The paper is grounded in a theory, implicit interaction framework, which describes two dimensions of implicit interaction: attentional demand and initiative. Attentional demand from the user can be placed in the background, for example, spell checking while typing, or in the foreground by underlining misspelled words. The system can take initiative and be proactive by identifying spelling mistakes while the user types or reactive by checking spelling only when the user has stopped typing.
I do not agree with some statements made by the authors in lessons learned. They state, “Implicit interactions often rely on the counterintuitive strategy of calling attention to observed or inferred perceptions”. This statement could mean that there is a mismatch between the user’s mental model and the system’s state. What is actually happening, in my opinion, is that the “inferred perceptions” are in fact different types of input that the user is not aware of. I would think that as these systems became more available to end users, they would be able to learn of these types of input. For example, most people before did not know how to use two fingers to zoom in and out of an interactive surface; now with Apple’s commercialization of the iTouch and variants, this is a ‘common’ gesture. What the real issue is the lack of understanding of the system’s input.
Another part of the sentence that I disagree with is the “strategy of calling attention.” All systems that take input have to provide feedback to the user. I thought we learned about this when building computational devices. To take an example in graphical user interfaces, feedback is provided when I type on a keyboard. The words are shown, blank circles are shown for password characters, or even a blinking caret is seen.
We have to learn how to educate people in learning the new types of input (the background of attentional demand), and how to provide adequate feedback to the user (the foreground). The paper provides an excellent example in a very narrow silo of the design space. There has been success in inferring implicit intention: the lights should go on when I enter a room using a motion sensor. People have learned to move around to trigger the sensor and keep the lights on; both the input and feedback mechanism are understood. An open research question is to have these inputs conform to some sort of norm, similar to how graphical user interfaces have norms for interaction (e.g., buttons, mouse icon, input caret).
