Web Tasking: An Investigation of End User Interaction for the Ideal Control Metaphor

Abstract

Users are finding multiple ways to utilize web applications (apps) outside of their typical self-contained purposes, resulting in an increasing need to connect apps together. This connectivity can be achieved through web tasking: the integration of web services/apps to achieve a personal goal. This research investigates the end user perspective, focused on comparing user interaction with web tasking interfaces through various analytical and empirical studies. These studies were divided into three distinct parts: i) Hierarchical Task Analysis (HTA) performed on existing web tasking interfaces as a usability benchmark, ii) creation and evaluation of a new interactive prototype, WebTasker, and iii) a full scale usability study with 16 participants evaluated four web tasking interfaces by performing 4 high complexity tasks and 4 low complexity tasks on 4 different interfaces (32 distinct tasks). A significant correlation was found between the number of keystrokes and mouse clicks/scrolls and task completion time; suggesting that simple task input counts could be used as an early usability predictor in web tasking interfaces. In addition, the HTA revealed several HF issues such as freedom of user actions by examining task structures (e.g. linear path versus wide HTA structure). The usability study showed that participants had poorer performance and found it more difficult to create web tasks with higher complexity. A mental model examination of composing web tasks found that participants preferred to enter task conditions first then actions. Web tasking is a new area in Human Computer Interaction (HCI) research and this research aimed to further develop web tasking interfaces to ultimately lead to an increase in user adoption of web tasking. The design of a new web tasking interface, WebTasker, utilized a journey line metaphor and proved to be successful in the usability study. It was recommended that it be further developed as a viable web tasking interface. Further lines or research are recommended including refining study tasks, dashboard development, and improvements to the WebTasker interface.