Dijkstra: On the cruelty of really teaching computer science (1988) [pdf]

Dijkstra: On the cruelty of really teaching computer science (1988) [pdf]
cs.utexas.edu

As almost all of Dijkstra's writing, this piece here is a plea for teaching formal methods. It's nicely written I admit, but it is also unbelievably smug and disconnected from reality. Even in physics or mathematics, which are obviously much more mathematically tractable than computer science, people do not do significant new things by using only formal methods. In fact, the technical achievement Dijkstra is most known for, Dijkstra's algorithm, was not invented using formal methods, but like most discoveries ever it occurred in a flash of insight after a long period of thinking, as he said in an interview. No matter how much one loves rigour, the dream of formalizing all reasoning has turned out to be a pipe-dream, as evidenced for example by the failure of the Hilbert program, and people everywhere proceed about their business as usual.

Mathematics is not, after all, just the manipulation of a string of symbols, and neither is Computer Science, even if all computers do is exactly this, because we humans are complex beings, largely driven by emotions, not consciously aware of all the reasoning processes going on, that are often reasoning in "fuzzy" ways (and frequently with good effects), as opposed to being machines executing chains of logical inferences. In fact, people are born with very different internal worlds, some people are really good at symbol manipulation, algebra, and language, their daily thinking consists of "talking to themselves", while some people naturally think by imagining pictures and having various loose sensual impressions, among the latter for example the great mathematician Poincare. Was Poincare in need of someone to "really teach him mathematics"? Unfortunately I think Dijkstra assumed everyone is identical to him.

Before preaching this too much, look at the practical implementations of his ideas on this, for example have a look at him lecturing on a technical topic:

http://www.youtube.com/watch?v=qNCAFcAbSTg

or read "A discipline of programming". As nice as the EWD prose is, I honestly can't stand the man talking about anything technical for more than 5 minutes. The more concrete the EWDs get the worse they are.

For some perspective on how science really gets done, as opposed to people's romantic images of how it is done, I recommend "The Sleepwalkers" by Arthur Koestler, about the Copernican revolution, and "The psychology of mathematical invention" by Hadamard.