Distracting mobile notifications are a high-profile problem but previous research suggests notification management tools are underused because of the barriers users face in relation to the perceived benefits. We posit that users might be more motivated to personalize if they could view contextual data for how personalizations would have impacted their recent notifications. We propose the ‘Reflective Spring Cleaning’ approach to support notification management through infrequent personalization with visualization of collected notification data. To simplify and contextualize key trends in a user’s notifications, we framed these visualizations within a novel who-what-when data abstraction. We evaluated it through a four-week longitudinal study: 21 participants logged their notifications before and after a personalization session that included suggestions for notification management contextualized against visualizations of their recent notifications. A debriefing interview described their new experience after two more weeks of logging. Our approach encouraged users to critically reflect on their notifications, which frequently inspired them to personalize and improved the experience of the majority.

Digital data, from texts to files and mobile applications, has become a pervasive component of our society. With seemingly unlimited storage in the cloud at their disposal, how do people approach data preservation, deciding what to keep and discard? We interviewed 23 participants with diverse backgrounds, asking them about their perceived digital data: what "stuff" they kept through the years, why, how they used it, and what they considered important. In an iterative analysis process, we uncovered a spectrum of tendencies that drive preservation strategies, with two extremes: hoarding (where participants accumulated large amounts of data, even if considered of little value) and minimalism (where they kept as little as possible, regularly cleaning their data). We contrast and compare the two extremes of the spectrum, characterize their nuanced nature, and discuss how our categorization compares to previously reported behaviors such as filing and piling, email cleaners and keepers. We conclude with broad implications for shaping technology.

An important visual cue for the distance to a target is its binocular depth, the disparity between the left and right eyes. We examined mid-air pointing on a large screen, varying the physical distances (depths) to targets. Welford's two-part formulation provided a better model than the one-part Fitts's Law formulation to predict movement time from movement amplitude and target width. Angular measures suggested by Kopper et al. did not improve the model. Consistent variations of Shoemaker et al.'s k-factor suggest target depth plays a role similar to gain for mid-air pointing. We compared both physical and virtual targets to determine if artificial binocular depth cues induce the same performance as purely physical binocular depth cues. Variation of the k-factor was different when virtual depth and physical depth were not identical. This has implications for calibrating 3-D virtual environments and for the design of interactive 3-D pointing techniques for those environments.