Illustrating Macros with Existing Documentation

This paper presents an approach that enlivens existing documentation, to efficiently support users performing procedural tasks. We describe an approach that automatically creates, in real-time, correspondences between the actions performed by users and the steps described in online documentation for the procedure being performed. Using these correspondences, our Macro Illustrator system can highlight the relevant portions of the documentation and provide the user with a visual indication of the progress being made. Users are therefore offered in-context documentation, which helps them track their current position in the procedure, and answers questions about the next steps to be taken. This approach works with existing documentation, requiring no additional markup on the part of the documentation author. We present an algorithm for extracting actions from documentation and aligning these extracted document actions with observed user actions through a system of similarity metrics. Empirical evaluation of this algorithm shows that it performs significantly better than a strawman approach. ACM Classification: I.2.7 [Artificial Intelligence]: Natural Language Processing Text Analysis. K.6.1 [Management of Computing and Information Systems]: Project and People Management Training. General Terms: Algorithms, Documentation, Human Factors