Replacing yourself with a very small shell script

March 26, 2012

One of the professors this semester gave me substantial liberty to design some of the homework assignments, liberty that I am now beginning to rethink.

The students have written programs implementing various alignment algorithms, which I grade automatically with a script that runs a bunch of unit tests and reports the results. Sometimes, though, the tests fail, and I must wade into the source to figure out what happened. If you ever want to ruin a nice weekend, try reading twenty different people’s perl code back to back.

I mentioned this to my fiancee’s father. His reply: “I’ve been using perl for twenty years, and I still learn something new every time I use it. That’s a problem.”


On a somewhat disappointing correspondence

March 24, 2012

I recall reading in Wittgenstein somewhere (Notes on Certainty?) of his fear of insanity: isn’t it possible, he asked, that one could be insane without knowing it? While Wittgenstein may have been right in point of fact to doubt his sanity (the state of his own mental health was often unclear), the general point seems underappreciated to me. We sometimes experience disagreements so bewildering that we may even be unable to locate the source of the disagreement. Such encounters leave us wondering whether the other parties to the epistemic conflict are really sane at all. By symmetry, shouldn’t we extend that doubt to ourselves?

(Attention conservation notice: long, boring, internet drama, arguably with no moral to the story)

Read the rest of this entry »

The Straussian impulse in homework

March 19, 2012

This semester I’m TA’ing for the upper level bioinformatics course.  As far as teaching assignments go, it’s really ideal: it’s my intended area of specialization, I get to work for wonderful professors and I’m afforded considerable latitude in teaching and assignment-writing.  (The latter two points were also true of cell bio last semester, though the first was definitely not.)  The bioinformatics course is unusual in that attracts both biology and computer science majors, roughly at a ratio of 1:2.  As such, there is considerable diversity in the students’ academic backgrounds, and it is expected that students will sometimes have knowledge of subjects (e.g. algorithms or biochemistry) that exceeds what is required for the course.

I think this may explain an annoying tendency in the graded assignments that I have noticed for some time but haven’t been able to put my finger on.

This tendency, which is probably more common among the CS kids than the biologists, is to lend obscure, esoteric interpretations to questions and answer them accordingly.  Mostly these involve pedantic or legalistic readings of questions which seem pretty facially straightforward to me.  For example, a warm-up question might ask for the output of an algorithm expressed in pseudo-code which prints the limiting ratio of F(n+1)/F(n), where F(n) is the nth Fibonacci number.  The facially straightforward answer to this question is 1.618… or phi, which most of the biologists were able to identify.  The legalistic, “gotcha” answer is 1, because division could be interpreted as C-style integer division.

But there is absolutely no reason to think that C-sytle semantics for the division operator were intended in the first place!  The behavior of the division operator in C was motivated by concerns of machine efficiency in general, and closure over ints specifically.  It’s not an especially important or universal fact about computer science, and there’s no reason to assume it’s implied in pseudo-code.  It actually requires “reading against the grain” of the assignment, which asked the student to identify the ratio in the next question.  Worse, the CS students know that they’re taking an interdisciplinary course where C’s division semantics is not common knowledge.

Nevertheless, the bureaucratic interpretation was depressingly common among CS students.  I would write it off as an isolated incident or a “thinko”, except that such mis-interpretations reccur with depressing regularity.  My pleas to adopt commonsensical readings of homework questions do not make a difference.

Why does this approach to homework bother me so much?  It assumes that there is an obvious interpretation of the assignment fit for commoners, and an esoteric interpretation of the text intended only for readers equipped with rare learning and subtle powers of discernment.  In other words, my students are Straussians about homework.