Publication Details

Download

Doris Hausen, Sebastian Boring, Saul Greenberg
The Unadorned Desk: Exploiting the Physical Space around a Display as an Input Canvas
In Proceedings of the 14th IFIP TC13 Conference on Human-Computer Interaction (INTERACT 2013). Cape Town, South Africa, September 2-6 2013.

In everyday office work, people smoothly use the space on their physical desks to work with documents of interest, and to keep tools and materials nearby for easy use. In contrast, the limited screen space of computer displays imposes interface constraints. Associated material is placed off-screen (i.e., temporarily hidden) and requires extra work to access (window switching, menu selection) or crowds and competes with the work area (e.g., as palettes and icons). This problem is worsened by the increasing popularity of small displays such as tablets and laptops. To mitigate this problem, we investigate how we can exploit an unadorned physical desk space as an additional input canvas. With minimal augmentation, our Unadorned Desk detects coarse hovering over and touching of discrete areas ('items') within a given area on an otherwise standard physical desk, which is used as input to the desktop computer. We hypothesize that peoples spatial memory will let them touch particular desk locations without looking. Consequently, and in contrast to other augmented desks, our system provides optional feedback of touches directly on the computers screen. To better understand how people make use of this new input space, we conducted two user studies: (1) placing and retrieving a varying number of items onto the desk, and (2) retrieving items from a predefined grid. We found that participants organize items in a grid-like fashion for easier access later on. When the given area is divided to represent a few large items, we (surprisingly) found no difference in error rates regardless of whether on-screen feedback was shown. Participants were also faster without feedback than with feedback. As the item number grew (and thus shrank in size to fit the area), participants increasingly relied on feedback to minimize errors, but at the cost of speed.