The author's goal is to systematically break down UNICOM's knowledge base in a manner that explains UNIX performance for users at different levels of expertise. That is, they want to determine what aspects of the knowledge base must be deleted to simulate intermediate and novice performance. They report on the efforts to simulate UNIX command performance acquisition at both the group level, and at the level of individuals. They model tasks whose solutions cannot be precompiled, because familiar elements are put together in a novel fashion. It is presumed that subjects are not recalling fixed scripts from memory. Rather, they are producing action plans for each task. They describe UNIX performance data, and the UNICOM model, and then describe how the model accounts for the failure of less expert users to produce effective action plans.<<ETX>>
[1]
John R. Anderson,et al.
Novice LISP Errors: Undetected Losses of Information from Working Memory
,
1985,
Hum. Comput. Interact..
[2]
Roberta L. Klatzky,et al.
Expertise in a Computer Operating System: Conceptualization and Performance
,
1990,
Hum. Comput. Interact..
[3]
Kurt VanLehn,et al.
Repair Theory: A Generative Theory of Bugs in Procedural Skills
,
1980,
Cogn. Sci..
[4]
Walter Kintsch,et al.
Toward a model of text comprehension and production.
,
1978
.
[5]
W. Kintsch.
The role of knowledge in discourse comprehension: a construction-integration model.
,
1988,
Psychological review.