Command verb

Abstract: A principle of good human-computer interaction design is to match the dialog to the way the user thinks about the task and the language. This dissertation studies language design based on developing an understanding of how the user mentally organizes the linguistic components of an interactive language. Two experiments were run to study the way expert users mentally organize the linguistic components of a user-initiated command language for an online computing system. In the first experiment subjects viewed command groups presented one word at a time and performed a written verbatim recall task at the end of each command group. The amount of time spent viewing each word was controlled by the user. From this temporal pattern of viewing times, inferences were made about how expert users organize a command in long term memory. Results indicated that a command is not treated as a unitary chunk, but is composed of understandable subchunks. In this case, the subchunks consisted of the command verb and its required positional parameters, and of keywords and their required keyword parameters. The second experiment was a replication of the first with subjects performing verbatim verbal recalls of the command stimuli. The verbal recalls were recorded and the temporal patterns analyzed for systematically large pauses indicating mental organizational boundaries. The analyses of both the recall and the viewing times support the findings of the first experiment. Cognitive psychology and psycholinguistic research illustrates that disruption of a person's mental organization results in errors for the task at hand. If we can design the dialog so that it does not disrupt the user's mental organization of the language, then we can reduce the overall number of user errors and improve productivity. This dissertation is a contribution to minimizing the likelihood of design errors due to the inattention and lack of understanding of the mental organization of the user.