October 11th, 2005

krazy koati

I'm very good at integral and differential calculus

``The exam consists of six questions,'' I explained, ``And you're to do your choice of five.'' Since I like giving my students some choice over the exam they take, though they usually end up skipping the easiest question.

``Any five?'' asked a student. I didn't quite know how to answer.

Alas, the exam went downhill from there. It started 15 minutes late -- another class was doing presentations, and they (of course) ran long, and it was the department chair running things so I didn't feel comfortable barging in and ordering everyone out.

More seriously, I horribly, outrageously, misjudged the length of the exam. It was supposed to be two hours, though I let students stay about as late as they want. It took just under four hours for everybody to finish off. That includes some time spent by one student, who ripped about 480 pages out of notebooks, and tried to find a stapler capable of adhering all this to the exam pages.

And worst of all: a question had a critical flaw. I took it from the book, in a section testing a particular algorithm, and put it on the exam without working it through. It turns out for that problem, the algorithm doesn't work. (Every numerical algorithm breaks for some problems.) Ow. I had pointed out repeatedly in class, while explaining the algorithm, that it might break and if it did, this is how you'd know -- but I didn't give any homework problems on which the algorithm breaks, and I didn't write in the question to consider the possibility that the algorithm might be inapplicable. My standing policy is, ``I do not ask trick questions,'' but this sure seems like one. I'm not sure how to grade it now.

In spam watch, I got, ``Re: Herbert Shinkle Pharmacy.'' I can believe there's a person named Herbert Shinkle, and he'd have few job options except to become a pharmacist or an actuary. Becoming both at once would just interfere with both careers, though.

Trivia: Eugene Cernan's share of the Life astronaut contract money, when he joined, came to US$16,250 per year. Source: The Last Man on the Moon, Eugene Cernan, Don Davis.

Currently Reading: Watt's Perfect Engine, Ben Marsden.