Don't know what a slide rule is for

[austin_dern]

Is this student sufficiently knowledgeable to skip my course? That's the question before me, and it's a good one. I've never before been asked whether a student might skip a course. You could aruge well that I'm not sufficiently knowledgeable to skip, much less teach, this course. The interview went fine, considering I lacked any not-stupid questions and the other instructors only asked follow-ups on mine. Administrators want me to have an opinion now.

The student offered an incredibly long and convoluted program he'd written, so I know he can write mind-choking code. And he wrote a small program showing one equation-solving algorithm. He gained my esteem by pointing out how the program was deficient -- in some cases it doesn't find solutions too accurately -- but lost some of that by making an incorrect guess about why. Decisions I'd be unqualified to make even if the data weren't ambiguous: that's what I'm here for.

Oh, and I spotted my first Singapore flag of the season today. I've mentioned before, it's illegal (or was until very recently) to display the flag except in the weeks around National Day; well, that's August 9. 'Tis the season.

Trivia: During his Mercury flight Gus Grissom's heartbeat reached a maximum of 171 beats per minute, and only dropped below 150 beats per minute for a few seconds of weightlessness. Source: ``Physiological Responses of the Astronaut in the MR-4 Spaceflight,'' C. Patrick Laughlin, and William S. Augerson, Results of the Second U.S. Manned Suborbital Space Flight, July 21, 1961.

Currently Reading: Profiles of the Future (1984), Arthur C Clarke.

Comments

[chipuni.livejournal.com]

I think that you've answered your own question. From the description you gave, the student wasn't sufficiently prepared to skip the course. He wasn't able to demonstrate his knowledge clearly, and he was missing a few of the fundamentals.

[austin-dern.livejournal.com]

Mmm ... but, he guessed wrong at the problem by guessing something trickier. He supposed the problem might be floating point glitches creeping up on the code, and worried about the numerical stability of the algorithm. (He didn't use the word, but that was the concern he was trying to express.) Actually he just set his tolerance for finding the zero higher than it needed to be, for a function that had a lot of zeroes close together.

It takes a certain sophistication -- and more than I think was needed for this course in previous terms -- to be aware of floating point and stability issues, so I'm not sure whether he took the wrong guess because he didn't know where to look, or thoughtlessly dismissed the simplest problem, or figured if I was going to listen to one possible reason he was going to present the ``smartest'' potential problem he could think of.

He did have the algorithm right, and written reasonably efficiently, and even commented it in pretty close to what I'd specify, suggesting to me he's at least in range.

[chefmongoose.livejournal.com]

Let's see. He can write big code. (did the long convoluted program work? Was it tested?) he can recognize a failing in his code in a smaller program.. but his code does have a failing. And though he knew about it, he didn;t know exactly why it failed.

Offhand I'd agree with Chip, except I'd need additional information. Is this a freshman-level course, or a higher level?

You could possibly give the student a test (say, last year's), and allow him to Test out rather than Skip out. That's the way Baylor worked it, and it seemed a fair system.

--Chairoscuro

[austin-dern.livejournal.com]

It's a freshman course -- basically, it's the class offered by the department so we have something with a lot of students in it -- intended to introduce programming and numerical methods. Nothing too deep.

His demonstrated humongous program worked, at least well enough to earn honors credit from a junior college, and he's got what look to be highest grades achievable for some courses and exams. (Singapore uses what looks like the British system of A-level and O-level and such exams, none of which make any sense to me.)

Exams from previous years would be great to use, except that I think the course has been focused too much on trivia in past years. For example, based on the homework assignments, last term spent at least four class-hours exclusively on the order of operations, which seems like overkill to me. Yes, you kind of need to know this, but is it worth testing the sample output of

int x = 3, y = 2, z = -2; x == y-x < y + z

...in this day and age? When, you know, you can use parentheses? Is it worth having multiple questions like this on the midterm and final? There's another question, one I can barely make out, that seems to test whether the student knows that one runs the executable, not the source code. Is this worth spending time on?

Give him his own program to fix.

[chahala.livejournal.com]

Hummm... I find that most modern science is based on knowing where nad how to look something up. So what you might do is have him correct the one big failure that you saw. Where he know there was a problem but not how to fix it.

He probable thinks that he is smart and bringing him into your class would just give you an unhappy student. So if you just hand him the test... here is your own program... you know there is a problem. Fix it. I will see you in two days.

That would give him a chance to prove himself to you and you can see if he is smart enough to look up the fix.

Good luck!

Chahala

Re: Give him his own program to fix.

[austin-dern.livejournal.com]

That's a good thought ... we did ask him why, if he was qualified to skip the course, he didn't take it anyway and get an easy A. His answer was that he wanted to learn a variety of topics, not redo what he's already learned. Granted that's the obvious answer, but a lot of people won't think to say that. I have to check when my decision's due.