Alyssa Nabors
Evin Schuchardt
Reference Information
Title: Speeding pointing in tiled widgets: understanding the effects of target expansion and misprediction
Authors: Jaime Ruiz, Edward Lank
Presentation Venue: IUI 2010: Proceedings of the 15th international conference on Intelligent user interfaces; February 7-10, 2009; Hong Kong, China
Summary
This paper discusses a technique to improve pointing with a mouse, trackball, stylus, etc. While the researchers list several possible techniques to improve performance, they focus on target expansion, which basically means that the target the person means to click on expands to make it easier to select. This isn’t as useful when the targets are in a dense arrangement, so the researchers also focus on a technique for predicting where a user will select in real time. A technique described in the paper is kinematic endpoint prediction (KEP), which “uses the motion profile of a user to define a region of interest on the display.” KEP is described in more detail in other papers referenced in this paper.
 |
| Image from paper: Unexpanded widget (top); expanded widget using displacement (bottom) |
The researchers perform two experiments. In the first experiment, they test to see if it is possible to expand a certain set of targets that are predicted to be selected and whether target displacement is acceptable. Target displacement refers to targets near the expanded targets shifting their positions to make room for the expanded target. They presented users with a random set of pointing tasks with some tasks using expansion with a simulated predictor said to be highly accurate and some using no expansion. There were seven possible tiles to select.
In the second experiment they used a similar procedure except they used the KEP algorithm to predict which targets should be expanded and tiled the entire screen with targets rather than just displaying seven targets.
Through their experiments they prove that target expansion does improve pointing and selecting performance if there is a small set of targets. They also find that whatever technique is used to predict where a user will select must have an accuracy greater than 56.5 percent. They also determined that there was a limit to how much a target should shift when expanded.
Discussion
Overall this was an interesting paper. Their experiments were thoroughly explained and their findings made sense. I wasn’t particularly interested in the topic. The researchers explain how improving pointing performance by even a small margin can greatly improve a user’s productivity with graphical user interfaces. At least when it comes to programs like Microsoft Word, I don’t think something like this would be useful. Having the most frequently used buttons along the top and using shortcuts seems sufficient to me.
Future work that the researchers mention include improving the accuracy of the KEP algorithm, using different expansion strategies and exploring other techniques to improve pointing performance.
My gut reaction is that I wouldn't want the target I'm trying to click on changing size of shape while I'm trying to click it.
ReplyDeleteThe said how effective the users were in hitting targets, but did they say if the test subjects felt this was something they would use?
No, I don't remember them providing any information on user opinion. They just focused on user accuracy with the system.
ReplyDeleteYou said, "improving pointing performance by even a small margin can greatly improve a user’s productivity..." I would be interested to see if this was true for both novice and experienced users of a particularly recognizable UI such as one like Word.
ReplyDeleteI agree with Josh. It's a novel concept, but how practical/useful would it really be?
ReplyDeleteMy guess is that this would be awesome every time it predicts what I was going to click on without me having to go through the effort, but the other half the time when it selects something that I didn't want would be enough for me to not use this. I like the research though, and think it could be improved in the future.
ReplyDelete