book: where the action is

I just finished possibly the best treatise on HCI I've read in a good long while -- Paul Dourish's Where the Action Is: The Foundations of Embodied Interaction. ... (yes, Amanda, it's even better than Hiroshi's)

Half HCI design and half philosophy class, Dourish's book attempts to unify recent research in tangible and social computing into a broader paradigm of "embodied interaction." Breaking with previous approaches steeped in Cartesian dualism and information processing cognitive psychology, embodied interaction is an extension of the phenomenology school of modern philosophy, as articulated by Heidegger, Merleau-Ponty, Wittgenstein, and others. It concerns engagement with the physical and social world, and recognizes that humans don't just follow set plans of action, but rather engage in a continuous improvisation in response to their surrounding contexts as they pursue goals and activities. Despite the strong philosophical slant, the book is very readable and in my opinion lays a nice foundation for both designing and interpreting interactive technologies. You won't find design guidelines or rules of thumb in this book, but you will find a useful perspective in which to frame your own analyses and ideas.

Having not read Heidegger before, I was particularly interested in the concepts of "ready-to-hand" (zuhanden) and "present-at-hand" (vorhanden) to describe our engaged status with an artifact. For example, when a skilled carpenter uses a hammer to drive a nail, the hammer is "ready-to-hand" -- the carpenter doesn't have to think about the fact that he is using a hammer, it cognitively becomes an extension of himself. That doesn't mean the carpenter can't also accidently smash his thumb, at which point he becomes keenly aware of the fact that he is wielding a hammer physically distinct from himself -- at this point the hammer becomes "present-at-hand." Apparently these concepts were introduced to the HCI community as early as 1986 by another book I need to read. Surprisingly, I haven't seen these concepts in the ubicomp literature, where they could help tease apart the oft-confused metaphors of "invisible" computing that are touted as ubicomp's ultimate goal.

The only thing I felt was noticably lacking from this book was a discussion of learning. Even embodied tasks often require some phase of learning, training, and familiarity gained through experience. I'd have loved to see an exposition of some of the features of such learning requirements as befits tangible and social computing applications. But as Dourish himself states, this book isn't intended to have all the answers. It is intended to be a beginning, to open a door into a new paradigm of interface research and development that "emphasizes the primacy of natural practice over abstract cognition." And that it most certainly does.