AN INTERACTIVE ENVIRONMENT FOR TOOL SELECTION, SPECIFICATION, AND COMPOSITION
Abstract
The application of computers to larger problem sets has led to a demand for integrated sets of tools that support a wide variety of functions while maintaining the appearance of a unified system. Subsequently, we are seeing the evolution of user environments with levels of sophistication that tax current software and hardware technologies. This thesis investigates many important issues pertinent to the design and implementation of such environments. In particular, we discuss the theoretical as well as practical aspects of developing an interactive user environment that supports an alternative to statement-at-a-time programming. A substantial part of this thesis is devoted to characterizing an interactive user interface based on menu-driven interaction. This characterization begins by describing the fundamental elements inherent to any menu-driven interface. It progresses through a sequence of successively more powerful models and culminates with the development of a hierarchical menu model that characterizes the full spectrum of properties intrinsic to menu-driven interaction. Additional consideration is given to user error recovery. We base user response reversal on the undo operation and provide a response sequence model to distinguish the characteristic behavior of the undo command from the undo response. This model is also used to describe several implementation alternatives. To complete user error recovery, we address several issues relevant to system and software failure. They include the time/space tradeoff between the complete rerun, and checkpointing strategies. To address some of the more practical issues, we discuss the design and implementation of OMNI: An Interactive Environment For Tool Specification, Selection, and Composition. Experimenting with several prototypes has led to the design of a dynamic environment that supports user defined characteristics as well as a method for minimizing menu network complexity while enhancing task specification capabilities. We also discuss several system performance issues that generalize to many user environments.
Degree
Ph.D.
Subject Area
Computer science
Off-Campus Purdue Users:
To access this dissertation, please log in to our
proxy server.