Abstract Olivier Danvy

The Exact Science of Programming


Programming computers used to be variously presented as an art (Knuth), a discipline (Dijkstra), a craft (Reynolds), and a science (Gries). Nowadays it is perceived as a mere skill.  Consequently, a program going wrong tends to be dismissed as a skill that was lacking.  However, the accompanying quip "nobody said that programming is an exact science" is absurd, as programming is perhaps the only science that is truly exact.

Besides justifying that programming is an exact science, the goal of this talk is two-fold:

(1) analyze why programming -- and computing in general -- has become such a disconsidered skill, and

(2) contribute to turning the tide and bringing science back into Computer Science.