Imaginary Interfaces
by: Sean Gustafson, Daniel Bierwirth and
Patrick Baudisch
Authors
These three researches are part of the Hasso Plattner Institute in
Potsdam, Germany, and are still doing research there today.
Presentation Venue
"UIST '10 Proceedings of the 23nd annual ACM symposium on
User interface software and technology" as per the document specs from the
ACM digital library, dl.acm.org. The presentation took place in New York
City, NY, 2010.
Summary
The general synopsis about this paper is that a movement to an imaginary interface would allow further miniaturization of mobile devices and one feasible starting place to begin work is creating a set of axes at any point using the user's non-dominant hand.
Hypothesis
Gustafson's work actually contained four hypotheses for different case studies which all led to his overall question about how far the human mind can be pushed to interact with an imaginary interface. He hypothesized that people with the imaginary interfaces would have fewer graffiti errors than reported by Ni and Baudisch, who provided earlier work of influence in the field. He also hypothesized that people would have an easier time using the same interface when they stand still between interactions as compared to when they turn around in between interactions. Coupled with this hypothesis was the hypothesis that users would have an easier time using the same interface when they use a hand to create the virtual space, versus not using a hand. The last hypothesis was that user error would increase as one moved further away from the coordinate axes provided by the non-dominant hand.
Methods
To test these hypotheses, Gustafson and company conducted a series of tests wherein different groups of people were chosen from their campus and given a set of tasks to perform with the imaginary interfaces. The cameras for the interfaces were not sufficient to capture the users' gesture data, so they had to do some adjusting to analyze their results. They reported this to be simple because most people paused between their interactions with the imaginary interface. The users had to draw various graffiti, draw and then turn to find the same point drawn, and find points in an imaginary grid extending from their non-dominant hand. Each of these three tests coincide with the four hypotheses in order, with the second and third hypotheses both belonging to the second test.
Results
In general, they believed their hypotheses to be somewhat true, but their studies lacked sufficient numbers to perform true statistical analyses. The users were able to draw simple objects, but increased strokes reduced accuracy. Rotation between interactions also reduced accuracy. However the third case showed some people better at finding coordinates and others worse.
Discussion
This is one of the few papers I've read that was actually interesting early on. It was like there was a little ink demon living in the pages waiting to throw out new blots of 'Oooh!' just when it looked like the Big Bad Reader was getting bored. In the context of the paper, the fact that they failed to capture the individual gestures with their cameras does not bother me. I still think the capture of hand maneuverability and accuracy holds the heart of their project - those beautiful electric signals spurting all over the human brain as it attempts to recreate what it has produced in subconscious memory upon a physical realm. Obviously, there needs to be more progress, but the starting point is profound in my opinion. If computers began with rigid lines and places to move your cursor before branching into a GUI with "windows" where the same basic principles apply but are expanded to a more generic use, then it is reasonable for me to assume that allowing a user to create their interface is the proper step towards eliminating the strict interface completely (which is the big step into the future that I look forward to). The lack of sufficient numbers to produce significant statistical findings takes some credibility away from the research, but I'm still inclined to enjoy the potential prospects from such a small sample space. On that same note, I do wonder what sort of skew the data will host for the fact that the test users were all in their twenties at a university. What happens to the grid test with someone who does not organize things in rectangles, but rather... dot categorization (akin to a dot population intensity map)? Still, the potential movement away from this glaring white screen I'm typing at is a prospect that moves me almost to tears (that might be the onion peels...).
