A schema for computer-human interface design.

The study advances the design of usable computer-human interfaces by proposing a schema which reduces the physical and mental effort expended during a terminal session, improves command consistency, and better meets user expectations. The schema initially proposes a task analysis which specifies software functionality in terms of the user's real-world perceptions and terminology. The schema then integrates the Object-Oriented paradigm and the GOMS model to isolate a set of generic methods which define the primitive sub-goals common to all object-oriented software designs. An algorithm is presented which utilizes these generic methods to guide the drafting of physical operator sequences that, when input, perform the atomic level sub-task revealed during task analysis. The algorithm next employs combinatorial techniques to identify multiple sets of physical operator sequences; each set being sufficient to generate all methods required for a useful interface. The schema references the generic methods a final time to generate the interface implicit within each physical operator sequence set. Two metrics are defined to identify the most usable of the useful interfaces. The mental effort metric, derived from Shannon's Information Theory, helps identify the useful interface which contains the most consistent methods. The physical effort metric, based on Fitts' Index of Difficulty, utilizes target size and distance to quantify the effort expended during terminal sessions. An experiment in which 42 subjects acquired triangular and rectangular targets provided data for validation of this metric. In the absence of an accepted standard for measuring target size and distance, the applicability of the Index of Difficulty to computer interfaces was investigated. This was accomplished without reference to questionable parameter values by comparing movement patterns between pairs of reflected cursor start locations. After finding the Index of Difficulty to be generally applicable, traditional methods of measuring target width and distance were tested against a measure based on assuming users minimize target acquisition effort. After showing the new measure to be preferred, the physical effort metric accepted.PhDUrban, Technological, and Environmental Planning; Sociotechnological PlanningUniversity of Michigan, Horace H. Rackham School of Graduate Studieshttp://deepblue.lib.umich.edu/bitstream/2027.42/105172/1/9635649.pdfDescription of 9635649.pdf : Restricted to UM users only