Over at NSLog(); the question of the day is how do you feel about open book tests. Ambivalent at best would be my first answer.
When I was in college, in the late 1970’s studying computer science, the only courses that regularly presented open book exams were the ones from my core curriculum, internal data structures, COBOL, online (CICS) programming, et-cetera. Sure, they could ask more obscure questions about the topic at hand, but if you were going to rely solely on the book or your notes to get all the answers then you weren’t going to do very well.
Without a solid basic understanding of the material at hand you’d lose too much time flipping through the book to complete the test. Most of the time I’d use the book to look up obscure tidbits, like the volume in bytes of a track on a 3350 disk drive.
The larger question for me is, “how do I feel about testing?” Throughout my high school and college careers I was constantly in trouble of failing or on academic probation. And yet I was one of the best programmers in my class and I have had a very successful working career for over 20 years now. Not only was I a poor test taker, I wasn’t motivated by the thought of getting a lessor grade.
I think that most testing fails because it is subjectively based on the teacher’s perspective of the material. If the test itself was a poor representation of the material covered in the classroom, then how could my failure on the test be an accurate measure of my knowledge of that material. Too often, in my experience as a student, the tests were wide of the mark when it came to accurately reflecting the course content.
Especially for the technical side of computer science, and any other discipline with a measurable outcome, I think a fairer evaluation would be of product produced and not just regurgitation of facts or theories.