The conclusion to my POPL 1998 talk

Concluding remarks in my POPL 1998 talk,
Parallel beta reduction is not elementary recursive

[joint work with Andrea Asperti, University of Bologna]

When the Chinese premier Zhou En-Lai was asked what he thought of the long-term importance of the French Revolution, he responded, "it is too early to tell."

Similarly, considering the future pragmatic implications of optimal evaluation as an implementation technique -- a different French Revolution, begun not by Maximilien Robespierre and Georges Danton, but by Gérard Berry and Jean-Jacques Levy -- I would like to say the same thing: it is too early to tell. But I have reasons for optimism.

(I apologize to Gérard for pairing him with Robespierre, but Jean-Jacques and Danton are a better match.)

I wish to emphasize, in summary, that the imperative principle of programming -- of which we heard Monday night, from one of France's most profound, respected, and eloquent philosophers of science [Gérard's invited talk, From Principles to Programming Languages] -- is a principle founded on the fundamentally royalist notions of command and state and assignment.

In marked contrast, the functional principle of programming -- championed by our modern-day Danton -- is founded on one of the fundamental Rights of Man: the freedom of expressions. And optimal evaluation is further based on the essentially populist principle of sharing work and resources: from each according to his computational abilities, to each according to his computational needs.

So if Danton could be with us today, he would not say, L'etat, c'est moi -- a royalist imperative. Instead, he would exclaim,

Je suis le peuple! Vive le lambda calcul! Et surtout, Vive la France!

Thank you.