What to Learn Next: Recommending Commands in a Feature-Rich Environment

Abstract

Despite an abundance of commands to make tasks easier to perform, the users of feature-rich applications, such as development environments and AutoCAD applications, use only a fraction of the commands available due to a lack of awareness of the existence of many commands. Earlier work has shown that command recommendation can improve the usage of a range of commands available within such applications. In this thesis, we address the command recommendation problem, in which, given the command usage history of a set of users, the objective is to predict a command that is likely useful for the user to learn. We investigate two approaches to address the problem. The first approach is built upon the hypothesis that users of feature-rich applications who have similar features tend to use the same commands, and also, a specific user tends to use commands with similar features. Building on this hypothesis, we describe a supervised learning framework that exploits features from a user-command network to predict new links among users and commands. The second approach is built upon three hypotheses. First, we hypothesize that in feature-rich applications there exists co-occurrence patterns between commands. Second, we hypothesize that users of feature-rich applications have prevalent discovery patterns. Finally, we hypothesize that users need different recommendations based on the time elapsed between their last activity and the time of recommendation. To generate recommendations, we obtain co-occurrence and discovery patterns from the command usage history of a large set of users of the same feature-rich application. Subsequently, for each user, we produce recommendations based on the user's command usage history, co-occurrence and discovery patterns, and time elapsed since the last command usage. We

Heard it through the Gitvine: an empirical study of tool diffusion across the npm ecosystem

“…Software engineering researchers and practitioners are often frustrated by the low adoption of practices and tools, such as static analysis, refactoring tools, program comprehension, or security testing, even though many of them are generally perceived to be beneficial. Researchers have found a wide range of technical and social pain points in tool adoption, such as false positives in static analysis tools [e.g., 39,64,71], missing trust in correctness [e.g., 74], crypting tool messages [e.g., 38,39], slow response times [e.g., 80], lack of workflow integration [e.g., 36,41,64,86], lack of collaboration support [39], lack of management buy-in [e.g., 18,80], overwhelming configuration effort [e.g., 25,36,39,71,80], and simply a lack of knowledge about tools [e.g., 60,74,87]. In response, most software engineering research has focused on technical solutions, such as improving functionality, accuracy, and performance [e.g., 8,64,72], improving usability [e.g., 38,46,51,72,79], and improving discoverability through recommendation mechanisms or process integration [e.g., 49,64,87].…”

IDE Interaction Support With Command Recommender Systems

“…Murphy-Hill et al [9] used six algorithms based on command popularity and collaborative filtering, by taking into account command executions histories. Zolaktaf and Murphy [12] suggested CoDis, which is based on command discovery patterns and co-occurrence of executions in sessions. Silva et al [11] performed static source code analysis to identify and rank opportunities for refactoring command executions.…”

“…• CoDis 6 : the algorithm that performed best in the offline study reported by Zolaktaf and Murphy [12]; and…”

“…On the other hand, only 35 of those commands were executed in the first three minutes. Because only one user received more than 6 such notifications, 12 in the period of 14 weeks, it is safe to assume that a higher frequency of notifications could lead to additional exploration of the IDE and even more executions of new (recommended) commands.…”

“…ICRS research community has designed, implemented, and tested several algorithms (see [7]- [12]), graphical user interfaces (GUIs) (see [13]- [15]), and even complete systems for recommending IDE commands in different settings (see [14]- [16]). However, these RSs have not been integrated in popular IDEs, yet; arguably because the accuracy and acceptance of the recommendations are still too low to enable practical usage.…”