ProjectProposal-JonathanChang
From CS160 User Interfaces Fa06
Contents |
Synchronized Testing Application
Problem Description
As students at Cal, we know how important it is for our professors to know what we think about our classes from the sheer number of surveys we are require to complete at the end of each semester. Such information is used to develop new curriculum, or adjust the existing lesson plan to teach in the most effective manner. As mentioned, there are surveys in place to evaluate the course itself and the various instructors, but it can never hurt to have more information.
Another issue is the presence of cheating during exams. At the undergraduate level, these practices are widespread and confound the efforts of both the professors and the university to accurately measure how much students have learned. Besides careful human monitoring and fallible pattern comparison during grading, there is no practical way of detecting copying during exams.
Target User Group
This application will most likely be used by Graduate Student Instructors and Teaching Assistants, who generally handle the particulars of grading and exam proctoring. Thus, the interface of the application should not interfere with the existing grading methods, nor add a considerable amount of time to their already hefty tasks. The data procured by this application will be of interest to professors for the design of the course, and perhaps to department or school administrators interested in how students are doing. For their benefit, this application should have a method of reporting data that is simple, perhaps even automated.
Problem Context and Forces
The exact form the application will take depend on what context it is being used in. Exams may be able to be tracked by the student filling it out, perhaps via their Student ID number at the top of each page. This would be adequate for collection of data like how long the student spends on each problem. For the function of cheating recognition, however, it is not necessarily the student's identity, but his/her geography that matters to the application. If two students are sitting next to each other, and their pen strokes are identical, barring handwriting differences, with a few seconds delay between them, there is likely some degree of collusion occuring. Behavior patterns like these must be studied, then programmed for recognition by the application while ignoring innocent behaviors.
Solution Sketch
