Location via proxy:   [ UP ]  
[Report a bug]   [Manage cookies]                
skip to main content
Automatic generation of context-sensitive animated help
Publisher:
  • George Washington University
  • Computing Facility School of Eng. & Applied Science Washington, DC
  • United States
Order Number:UMI Order No. GAX92-00864
Reflects downloads up to 23 Sep 2024Bibliometrics
Skip Abstract Section
Abstract

Context-sensitive animated help uses simple 2-dimensional animation to demonstrate how to perform a computer application task within the context in which help is requested. To automatically generate such help, descriptive information about an application, its interface design, and runtime context must be available to a help generation system. This dissertation addresses how to represent domain-independent help knowledge, which is suitable for automatic generation of procedural animated help. The dissertation also presents the design of a help architecture which allows help and user interface components to share runtime contextual information. Within the environment this architecture supports, help can be tailored specifically to the user's current context.In this dissertation research, a knowledge representation model is developed that sufficiently details an application interface for both animated help generation and user interface control purposes. The knowledge model captures the semantics of application actions, their relationships to user interface components and functionality, and interface details required to support accurate demonstrations that are adaptable to current contexts. A model for a runtime support architecture with an integrated help component is also developed, in which help and user interface components share a common knowledge base to generate context-sensitive animated help automatically at runtime and to assist the management of use interfaces. The animated help component has access to contextual information, and employs artificial intelligence planning techniques to tailor animation scripts specifically to current contexts. Based on these models, a system called Cartoonist is implemented to demonstrate the usefulness of both the knowledge representation and the software architecture models.

Contributors
  • Georgia Institute of Technology

Recommendations