Thursday, February 26, 2009

An Exploration of Pen Rolling for Pen-based Interaction

By: Xiaojun Bi, Tomer Moscovish, Gonzalo Ramos, Ravin Balakrishnan, Ken Hinckley


The study explored the possibility of augmenting the pen's usability by introducing input from rolling it around its longitudinal axis. Two experiments were carried out to determine the feasibility and the possible applications. One experiment distinguished the intentional pen rolling from incidental, and the other experiment explored the parameter range at which users would be able to use it for application purposes.

Experiment 1: Incidental vs. Intentional Pen Rolling

The first experiment was on how the pen rotated under normal usage such as writing and drawing. Four tasks were given to participants to monitor pen movement:

1. Free Drawing
2. Writing
3. Screen Tracing
4. Table Tracing

Photos from "An Exploration of pen rolling for Pen-based Interactions", By Bi

Each task required a different type of pen movement and was monitored with cameras positioned around the setup. The cameras focused on markers that were attached to the pen. The parameters defined in this experiment were rolling angle and rolling speed. The experiment was carried out by setting the pen at 0, 45, and 90 degrees, and then measure from there the distance of movement as well as speed.

Rolling speed was fastest in writing and slowest in free drawing.
Rolling angle increases with the distance from the starting point of a stroke. There is a tendency for clockwise rotation. The mean rolling angle was 3 degrees. The mean rolling speed was 20 degrees/sec. To classify as an incidental roll, the rolling angle must fall within the [-10, 10] degree range and the rolling speed be in [-30, 30] degree/s range. If it falls outside of that range, it will considered intentional rolling.

Experiment 2: Intentional Rolling
The second experiment looked at the parameters during intentional rolling. The tasks given to participants included having them move a pen into a certain range of degrees. Variables that were monitored were the angular distance to the target and the angular width of the target sector. These variables represented the "usable range" and the "easily discriminable rotation".

Easily discriminable range was set at more than 10 degrees. As the angular target width decreased, so did the user's performance. (i.e. more errors, overshot target sector, and took more time.) Usable range was determined to be within [-90,90] degrees. The users' performance deteriorated as the range increased beyond that.

Pen Rolling Based Applications and Interactions
Users were subjected to applications such as using rolling to select a tool or to scroll down a page. Visual experiments were carried out where a user had to direct a puzzle piece using a pen and rotate the piece by rolling the pen. Pen-rolling could be used in parallel with other inputs. Another application involved users having to roll the pen a certain number of degrees to select a tool in a paint-like application. tools were set within certain ranges of degrees.

Photos from "An Exploration of pen rolling for Pen-based Interactions", By Bi

Participants who used the applications were generally able to quickly get use to it. Most felt the preset parameters were easy enough to use and some even suggested a bigger roll angle. One complaint that was raised was the difficulty to take advantage of the multiple input ability of the pen. It was difficult for the user to control direction, position and magnitude simultaneously.

The paper concluded that, while this is feasible and the parameters set are reasonable, additional research and evaluations need to be conducted and that pen-rolling based applications can indeed be beneficial to users. Additionally, it might be a good idea to let users customize the thresholds such as rolling angle.


  1. It would be interesting to see this technology when it is more developed. The idea of being able to change different settings based on pen rolling is a novel concept, but only if it works. This would certainly save graphic designers a lot of time being able to access the different things they needed without moving there pen far from where they were. Seems like it is still a ways off before the technology can be made mainstream though.

  2. That is incredible to think we can do stuff like this. Some firefox addons have stuff like this where you can drag a line and based on what you do, it will refresh, go back, etc.

  3. This is an awesome idea, they need to try and apply this to mice. I don't (most don't) use a pen with their computers, but we all use mice. If I could spin my mouse in place and have that actually do something I would like that. Spinning the mouse might be a little less natural, but the rewards would be worth the awkwardness.

  4. Good job with the presentation. It's no wonder we both chose this topic. I certainly found it to be the most interesting of the UIST 2008 papers.