Teaching

For some reason, since the weekend I'm pretty down and yesterday's events did improve anything. I had an exercise session in Software Verification and the exercises were to be done with pen and paper which suited me fine since I don't know much about separation logic and I am curious to see how convenient it is to use it to solve day to day programming problems. First disappointment: instead of presenting a collection of usable manipulation rules and theorems, he gave us an intuitive notions of what the logic is about and expected that it would be enough to solve the problem. This defeats the whole purpose of using formal logic and I equate with no less than giving up the teaching of the topic. To my liking, something good enough for the exam but unusable afterwards is a total failure of teaching Second disappointment: we had to write a program with a loop and prove its partial correctness. He mentioned techniques of weakening the postcondition to obtain a loop invariant but he said that he would not teach it since they were not likely to have to find an invariant at the exam. Instead, he suggested to (with no alternative, we could say imposed on) the students to execute their loop on paper and, through guesswork, to find the relation that is preserved throughout the iterations. This is at best a very clumsy way of finding a loop invariant and I am sure that, for many students, it has driven home the message that formal methods is no more than an academic exercise instead of something useful that they can use when dealing with hard problems. And it is not like he did not know that they are very useful for the direct derivation of programs or that he thinks that it is inefficient problem solving. The only conclusion left is that he thinks students are not up to it. That's sad to have such a poor esteem of the people that came all that way to learn something. And I don't mean that I'm the only one victim of that practice (actually, I don't consider myself a victim at all since I think lectures are only for administrative purposes; my true learning is done on my own time with books and smart people) but I think of the people that worked all through their bachelors to end up with some farce of a lecture. Even though I am aware that not everybody is as motivated as me for learning new material, I would think that the time spent there could be better spent doing something else. I feel better already, having said that! Simon

Pondering about Reasoning Errors

This note is mostly about teaching and, in general, communicating with other human beings. I have thought about writing it for a while but I was afraid not to have enough material. I'm writing it now because of new observations and I don't care too much about the length, today. In logic, we have the two important notions of soundness and of completeness. In short, a logical system is sound if you can only use it to prove statements that are true (but maybe not all of them) and complete if you can prove all statements that are true (but maybe also some false statements). I mention them because they appear more and more relevant to the way I choose to explain or with which I choose to help people. In such cases, it is necessary to make assumptions in order to explain or skip notions. Two kinds of errors are of interest for me here: either I assume that the other person does not know something and choose to explain it while he or she already know about it or I assume that he or she does and skip it while he or she does not know about it. In the first case, the reaction can be to be insulted because the attitude seems patronizing because the notion seems obvious. In the second case, the other person can miss a notion and be too afraid to appear stupid to ask about it. While I was an intern at A2M (a game developer in Montreal), my supervisor, which I regard with high respect, taught me that it is desirable to make communications as simple and concise as possible and to trust that the other person (in this situation an engineer or a scientist) will not be shy to ask about the missing parts. During the last hour, I've been thinking about this post while doing my groceries and I realized that there are a handful of people that I trust enough to be very straightforward in my communications because I've been talking to them for a long enough period to see that they are active while listening. For the others, it does not mean they are any less. It just means that I don't know if they will ask questions and, from time to time, I don't want to risk them misunderstanding me. In no case is this a sign that I mistrust someone's intelligence because I noticed recently that I was too explicit even with some people which I have in high esteem. This is pretty short but it's all I could summon before my stomach started threatening me. Simon August 14th 2009 Zürich