|
|
|
-- Kitajima (1996)Kitajima, M. (1996). Model-based analysis of required knowledge for successful interaction with a novel display. ICS-TR, 96-03. Institute of Cognitive Science, University of Colorado.
Model-based analysis of required knowledge for successful interaction with a novel displayThe goal of this study is to identify the knowledge users need to discover correct actions in novel interfaces. Users' task performances are simulated with the Kitajima and Polson (1995) comprehension-based model of action planning for display-based human-computer interaction (HCI). The action planning model is provided with a sufficient set of interaction knowledge for performing a graphing task, and then simulates the task in various situations defined by the combination of the following two conditions. The first condition concerns the degree of specificity of goal representations. The model is provided with goals with varied levels of detail, including the representations of task goal and desired device state. The second condition concerns the number of screen objects that the model considers in the action selection processes. The model pays attention to either a limited number of task relevant screen objects or all the screen objects from the entire screen. Results demonstrate that specific representations of a desired device state are critical for correct action planning. However, even without these representations, which is likely in the use of novel interfaces, the action planning model can act correctly when a task goal representation closely matches the correct screen object's label and when the model focuses on the task relevant area of the screen.
Download
|