Improving keystroke savings is a long-term goal of text input research. We present a study into the design space of an abbreviated style of text input called C-PAK (Correcting and completing variable-length Prefix-based Abbreviated Keystrokes) for text entry on mobile devices. Given a variable length and potentially inaccurate input string (e.g., “li g t m”), C-PAK aims to expand it into a complete phrase (e.g., “looks good to me”). We develop a C-PAK prototype keyboard, PhraseWriter , based on a current state-of-the-art mobile keyboard consisting of 1.3 million n -grams and 164,000 words. Using computational simulations on a large dataset of realistic input text, we found that, in comparison to conventional single-word suggestions, PhraseWriter improves the maximum keystroke savings rate by 6.7% (from 46.3% to 49.4,), reduces the word error rate by 14.7%, and is particularly advantageous for common phrases. We conducted a lab study of novice user behavior and performance which found that users could quickly utilize the C-PAK style abbreviations implemented in PhraseWriter, achieving a higher keystroke savings rate than forward suggestions (25% vs. 16%). Furthermore, they intuitively and successfully abbreviated more with common phrases. However, users had a lower overall text entry rate due to their limited experience with the system (28.5 words per minute vs. 37.7). We outline future technical directions to improve C-PAK over the PhraseWriter baseline, and further opportunities to study the perceptual, cognitive, and physical action trade-offs that underlie the learning curve of C-PAK systems.

The effective use of assistive interfaces (i.e. those that offer suggestions or reform the user’s input to match inferred intentions) depends on users making good decisions about whether and when to engage or ignore assistive features. However, prior work from economics and psychology shows systematic decision-making biases in which people overreact to low probability events and underreact to high probability events – modelled using a probability weighting function. We examine the theoretical implications of this probability weighting for interaction, including its suggestion that users will overuse inaccurate interface assistance and underuse accurate assistance. We then conduct a new analysis of data from a previously published study, quantifying the degree of bias users exhibited, and demonstrating conformance with these predictions. We discuss implications for design, including strategies that could be used to mitigate the deleterious effects of the observed biases.

Studies in psychology have shown that framing effects, where the positive or negative attributes of logically equivalent choices are emphasised, influence people's decisions. When outcomes are uncertain, framing effects also induce patterns of choice reversal, where decisions tend to be risk averse when gains are emphasised and risk seeking when losses are emphasised. Studies of these effects typically use potent framing stimuli, such as the mortality of people suffering from diseases or personal financial standing. We examine whether these effects arise in users' decisions about interface features, which typically have less visceral consequences, using a crowd-sourced study based on snap-to-grid drag-and-drop tasks (n = 842). The study examined several framing conditions: those similar to prior psychological research, and those similar to typical interaction choices (enabling/disabling features). Results indicate that attribute framing strongly influences users' decisions, that these decisions conform to patterns of risk seeking for losses, and that patterns of choice reversal occur.

Many interfaces attempt to assist users by offering suggestions, predicting actions, or correcting user input. Although these attempts may often be beneficial, they will occasionally fail, and prior research from psychology and behavioral economics on loss aversion suggests that such failures may have an overweighted impact on subjective preferences. To better understand the relationship between interface outcomes and subjective preferences, we adapted to interaction a model from behavioral economics describing reference-dependent preferences. Two experiments examined our model’s predictions: both involved word-snapping interface assistance during text selection, with subjects choosing whether they preferred unassisted text selection (no snapping) or assisted text selection (snapping). In Experiment 1, the word-snapping feature could be disabled when it was unhelpful by backtracking the selection, which required progress losses in terms of target characters selected. In Experiment 2, word-snapping could be disabled by pressing a modifier key and waiting for an animation to complete, without need for the loss of target characters. Time and error performance with both experimental methods were comparable. The model predicts an aversion to progress losses in Experiment 1 that should be neutralized in Experiment 2, and the results of both experiments conformed to these predictions. We discuss the implications of the model as a platform for understanding user preferences more broadly, especially when considering interfaces that risk periodically failing to meet user expectations.

Finger pressure offers a new dimension for touch interaction, where input is defined by its spatial position and orthogonal force. However, the limited availability and complexity of integrated force-sensing hardware in mobile devices is a barrier to exploring this design space. This paper presents a synthesis of two features in recent mobile devices - a barometric sensor (pressure altimeter) and ingress protection - to sense a user's touch force. When a user applies force to a device's display, it flexes inward and causes an increase in atmospheric pressure within the sealed chassis. This increase in pressure can be sensed by the device's internal barometer. However, this change is uncontrolled and requires a calibration model to map atmospheric pressure to touch force. This paper derives such a model and demonstrates its viability on four commercially-available devices (including two with dedicated force sensors). The results show this method is sensitive to forces of less than 1 N, and is comparable to dedicated force sensors.

Active Edge is a feature of Google Pixel 2 smartphone devices that creates a force-sensitive interaction surface along their sides, allowing users to perform gestures by holding and squeezing their device. Supported by strain gauge elements adhered to the inner sidewalls of the device chassis, these gestures can be more natural and ergonomic than on-screen (touch) counterparts. Developing these interactions is an integration of several components: (1) an insight and understanding of the user experiences that benefit from squeeze gestures; (2) hardware with the sensitivity and reliability to sense a user's squeeze in any operating environment; (3) a gesture design that discriminates intentional squeezes from innocuous handling; and (4) an interaction design to promote a discoverable and satisfying user experience. This paper describes the design and evaluation of Active Edge in these areas as part of the product's development and engineering.

Word–Gesture keyboards allow users to enter text using continuous input strokes (also known as gesture typing or shape writing). We developed a production model of gesture typing input based on a human motor control theory of optimal control (specifically, modeling human drawing movements as a minimization of jerk—the third derivative of position). In contrast to existing models, which consider gestural input as a series of concatenated aiming movements and predict a user’s time performance, this descriptive theory of human motor control predicts the shapes and trajectories that users will draw. The theory is supported by an analysis of user-produced gestures that found qualitative and quantitative agreement between the shapes users drew and the minimum jerk theory of motor control. Furthermore, by using a small number of statistical via-points whose distributions reflect the sensorimotor noise and speed–accuracy trade-off in gesture typing, we developed a model of gesture production that can predict realistic gesture trajectories for arbitrary text input tasks. The model accurately reflects features in the figural shapes and dynamics observed from users and can be used to improve the design and evaluation of gestural input systems.

User interfaces often attempt to assist users by automating elements of interaction, but these attempts will periodically fail -- impeding user performance. To understand the design implications of correct and incorrect assistance, we conducted an experiment in which subjects selected their preferred of two interfaces (neutral and snapping) for a series of 10 drag-and-drop tasks. With neutral the dragged object moved pixel-by-pixel, and with snapping the object snapped to a grid. Snapping trials were engineered to provide controlled levels of objective performance gains and losses with respect to neutral: gains were achieved when the target was aligned with the grid, and losses were achieved through misalignment -- which required subjects to drop the object, hold a key, and complete the task using a finer movement resolution. Results showed a significant preference for the snapping interface, even when losses impaired performance.

Mobile keyboards often present error corrections and word completions (suggestions) as candidates for anticipated user input. However, these suggestions are not cognitively free: they require users to attend, evaluate, and act upon them. To understand this trade-off between suggestion savings and interaction costs, we conducted a text transcription experiment that controlled interface assertiveness: the tendency for an interface to present itself. Suggestions were either always present (extraverted), never present (introverted), or gated by a probability threshold (ambiverted). Results showed that although increasing the assertiveness of suggestions reduced the number of keyboard actions to enter text and was subjectively preferred, the costs of attending to and using the suggestions impaired average time performance.

Psychological research has shown that 'peak-end' effects influence people's retrospective evaluation of hedonic and affective experience. Rather than objectively reviewing the total amount of pleasure or pain during an experience, people's evaluation is shaped by the most intense moment (the peak) and the final moment (end). We describe an experiment demonstrating that peak-end effects can influence a user's preference for interaction sequences that are objectively identical in their overall requirements. Participants were asked to choose which of two interactive sequences of five pages they preferred. Both sequences required setting a total of 25 sliders to target values, and differed only in the distribution of the sliders across the five pages -- with one sequence intended to induce positive peak-end effects, the other negative. The study found that manipulating only the peak or the end of the series did not significantly change preference, but that a combined manipulation of both peak and end did lead to significant differences in preference, even though all series had the same overall effort.

Edge-scrolling allows users to scroll a viewport while simultaneously dragging near or beyond a window's edge. Common implementations rely on rate control, mapping the distance between the pointer and the edge of the viewport to the scrolling velocity. While ubiquitous in operating systems, edge-scrolling has received little attention, even though previous works suggest that (1) rate control may be suboptimal for isotonic pointing devices like mice and trackpads and (2) space beyond the window's edge might be scarce, limiting scrolling control. To address these problems, we developed Push-edge scrolling (and Slide-edge scrolling, its inertial variant), two novel position-based techniques that allow scrolling by "pushing" against the viewport edge. A controlled experiment shows that our techniques reduce overshoots and offer performance improvements by up to 13% over traditional edge-scrolling.

Touch scrolling systems use a transfer function to transform gestures on a touch sensitive surface into scrolling output. The design of these transfer functions is complex as they must facilitate precise direct manipulation of the underlying content as well as rapid scrolling through large datasets. However, researchers' ability to refine them is impaired by: (1) limited understanding of how users express scrolling intentions through touch gestures; (2) lack of knowledge on proprietary transfer functions, causing researchers to evaluate techniques that may misrepresent the state of the art; and (3) a lack of tools for examining existing transfer functions. To address these limitations, we examine how users express scrolling intentions in a human factors experiment; we describe methods to reverse engineer existing `black box' transfer functions, including use of an accurate robotic arm; and we use the methods to expose the functions of Apple iOS and Google Android, releasing data tables and software to assist replication. We discuss how this new understanding can improve experimental rigour and assist iterative improvement of touch scrolling.

We describe a method for applying gain to events reported by scrolling input devices such as scroll wheels. By treating document length as an input to our gain functions, the method allows rapid document traversal regardless of document length; it also allows slow and precise scroll control at shorter distances. An initial experiment characterises four diverse scrolling input devices -- a standard 'notched' scroll wheel, a high performance 'inertial' wheel, an isometric scrolling joystick, and a trackpad -- and the results are used to calibrate several gain function parameters. A second experiment validates the method, showing that it allows faster scrolling in long and short documents than current scrolling-device gain methods, and that subjective preferences favour it.

Scrolling is controlled through many forms of input devices, such as mouse wheels, trackpad gestures, arrow keys, and joysticks. Performance with these devices can be adjusted by introducing variable transfer functions to alter the range of expressible speed, precision, and sensitivity. However, existing transfer functions are typically "black boxes" bundled into proprietary operating systems and drivers. This presents three problems for researchers: (1) a lack of knowledge about the current state of the field; (2) a difficulty in replicating research that uses scrolling devices; and (3) a potential experimental confound when evaluating scrolling devices and techniques. These three problems are caused by gaps in researchers' knowledge about what device and movement factors are important for scrolling transfer functions, and about how existing devices and drivers use these factors. We fill these knowledge gaps with a framework of transfer function factors for scrolling, and a method for analysing proprietary transfer functions---demonstrating how state of the art commercial devices accommodate some of the human control phenomena observed in prior studies.

Several enhanced pointing techniques aim to reduce the Fitts' law targeting distance by providing multiple target trajectories in the hope that a shorter path is available. However, these techniques introduce a search or decision component to pointing users must examine the alternatives available and decide upon the trajectory to use. We analyse these difficulties, present a methodology for examining them as well as other behaviour issues, and report empirical results of performance with pointer wrapping and Ninja cursors. Results show that offering multiple trajectories incurs a significant search or decision cost, and that users are therefore poor at capitalising on the theoretical benefits of reduced target distance.

Interface guidelines encourage designers to include shortcut mechanisms that enable high levels of expert performance, but prior research has demonstrated that few users switch to using them. To help understand how interfaces can better support a transition to expert performance we develop a framework of the interface and human factors influencing expertise development. We then present a system called Blur that addresses three main problems in promoting the transition: prompting an initial switch to expert techniques, minimising the performance dip arising from the switch, and enabling a high performance ceiling. Blur observes the user's interaction with unaltered desktop applications and uses calm notification to support learning and promote awareness of an alternative hot command interface. An empirical study validates Blur's design, showing that users make an early and sustained switch to hot commands, and that doing so improves their performance and satisfaction.

It is easy for users to lose awareness of their location and orientation when navigating large information spaces. Providing landmarks is one common technique that helps users remain oriented, alleviating the mental workload and reducing the number of redundant interactions. But how many landmarks should be displayed? We conducted an empirical evaluation of several relationships between the number of potential landmarked items in the display and the number of landmarks rendered at any one time, with results strongly favouring a logarithmic relationship.