Paper Reading #9: The Satellite Cursor: Achieving MAGIC Pointing Without Gaze Tracking Using Multiple Cursors

Comments
Jimmy Ho
Derek Landini

Reference Information
Title: The Satellite Cursor: Achieving MAGIC Pointing Without Gaze Tracking Using Multiple Cursors
Authors: Chun Yu, Yuanchun Shi, Ravin Balakrishnan, Xiangliang Meng, Yue Suo, Mingming Fan, Yongqiang Qin
When/Where: UIST 2010, New York City, New York



Summary
This article discussed the satellite cursor, a technique of using a cursor where there are multiple cursors, one associated with every target on the screen. Each cursor moves synchronously. The idea is that by having multiple cursors, each associated with a specific target, pointing performance will be improved by reducing the amount of movement required for input. The cursor's position is calculated such that the user can only select one target at a time.

An example of Satellite Cursor, with one cursor per green target, all moving synchronously. Source: Article.

The satellite cursor uses an "Aggregate and Expand" algorithm for calculating positions of all of the satellite cursors. This algorithm aggregates all targets to the main cursor then expands all image targets to the same position of their targets. The use of empty space is considered, and by aggregating the targets to be as compact as possible, all empty space is eliminated, reducing the distance the mouse needs to move. The only constraint for this algorithm is that there should be no overlap between targets.

Different methods of target layout are considered, and user testing is done to test the advantages and disadvantages between some of them. Of two studies, each consisted of asking participants to select targets on the screen. The first one was simpler, and measured completion time. Results concluded significant differences between layouts, and demonstrated that the performance of the cursor can be modeled well by Fitts' Law. The second study also measured completion time, but this time with different cursor layouts and selection strategies.

Discussion
This paper was difficult to get through. While it explained product design and testing, much of the explanation involved numbers and equations, manipulating Fitts' Law, which I have only briefly read about before. Due to my unfamiliarity with this, it made it difficult to understand quite what was being discussed.

That being said, it was fairly interesting to read about a new concept for cursors. The idea of having a cursor associated with each target, each moving synchronously with the main mouse sounds very confusing, but could also be very useful in eliminating time wasted dragging a single cursor back and forth across the screen.