Demonstration-based help: a case study

Abstract

The usability of today’s applications is of utmost importance, and to fully achieve it some changes are yet to be adopted. Help systems are part of the applications and allow users to understand it and use it. However, this is one of the fields that some corporates still neglect. Moreover, every system has its characteristics and customizations and being able to explain how something works on a generic but efficient way is a major challenge. Studies have been conducted and applications, especially on a daily-use context, are still cause of frustration to its users. Contextual and demonstration-based strategies are examples of beneficial approaches to improve the status. Additionally, current technologies like picture-driven computing and automation are enablers facilitating the interaction with “closed” applications, whose source codes are not available, and allowing for the process to be reproduced automatically on any platform. Our solution deploys fast easy learning. It allows users to learn how to perform tasks by watching them get done on their own system. This was done by the manual creation of picture-driven scripts with the use of a tool named Sikuli. This paper presents our tool and a preliminary case-study. A preliminary evaluation with users was made and positive results were shown. When tracking metrics relative to a first interaction with a system/realization of a specific tasks, users achieved the goal five times faster with the demonstration-based help tool. On second time executions, the performance from users that used our tool and users that learned through exploration and manual readings were similar, indicating that learning by observation does not detract the learning process. The solution is now to be implemented in a technology corporation where problematic tasks and platforms will be identified, automation scripts developed, and an evaluation with its employees will be conducted. From this evaluation, the usability of the tool will be measured, as learning improvements and increased task performance will be tracked. The availability to help other users, developing and sharing scripts is also one of the research points.