Automatic Generation of Context-Sensitive "Show and Tell" Help

Abstract

Textual help has become insufficient to guide users to carry out procedural tasks. Help for a direct manipulation interface requires that textual help be verbose in order to describe procedures accurately and specifically. In this paper, we attempt to alleviate this problem by presenting procedural help using coordinated textual and animated help. Our help presentation is dynamically created at runtime and is derived from the underlying user interface representations of the application model. Generating help from the user interface presentations allow help contents to be tailored specifically to the current context.