Paper Reading #16: The Satellite Cursor

Comments:

Adam Friedli

Shena Hoffmann

Reference Information:

Title: The Satellite Cursor: Achieving MAGIC Pointing without Gaze Tracking using Multiple Cursors

Authors: C. Yu, Y. Shi, R. Balakrishnan, X. Meng, Y. Suo, M. Fan, Y. Qin.

Presentation: UIST' 10, October 3-6, 2010, New York, New York.

Summary:

The Satellite Cursor is a technique developed for use of multiple cursors with the goal of improving pointing performance. The developers are able to achieve their goal by reducing input movement, how much movement the user needs do achieve before reaching the target. Previous techniques were based on Fitts Law, they tried to improve pointing performance my reducing the amplitude and width of the targets and motor space. However, these techniques do not focus on reducing distraction factors caused  by bypassed targets.

The Satellite Cursor focuses on achieving an appropriate layout of targets in motor space as well as minimum distraction. It employs one cursor per target, but only one cursor is actually on the target when the user is selecting it. In other words, there is only one cursor able to select a target at a time. Thus, there is one constrain that must be met, targets cannot overlap. When the users moves the mouse, all cursors move synchronously. The developers propose a two step algorithm, "Aggregate and Expand." In the 'Aggregate' step all targets are aggregated to the main cursor. Then in the 'Expand' step, the location of all satellite cursors are calculated in order to distribute the targets to the satellite cursors. 

In this image, there are four different satellite cursors that move synchronously, and demonstrates how only one cursors is able to select a target at a time.

In order to evaluate the Satellite Cursor, developers carried out two experiments, one was a simple pointing task, while the second one was a more complex task with multiple targets of varying layout densities. Based on their results, there are two main areas where the Satellite Cursor are successful: it can save significant mouse movement to reach a target, and it is especially beneficial for target layouts that are sparse. They concluded that the satellite cursor performance can be modeled using Fitts Law successfully.

Discussion: Even though I think this is a very creative development, and the results show that it should decrease the mouse movement to reach a target, I am not sure how effective this would be. I mean, having all these cursors floating around could cause more distraction than bypass targets. I cannot tell for sure, since I have never tried something like this before. They do discuss in the paper how clutter can affect visual aspect of it, that is why they affirm it is more effective in sparse layouts. I would like to try it out and see how confusing or not confusing it gets.