October 1996

Root Page of Article: Coming Out of the Closedt World, by John Horberg

Decent if Sparse Overview

Admittedly, all this can be found elsewhere, in much greater detail and more complete form. Still, Edwards' context gives a new perspective that seems worthwhile to me.

For example, Edwards notes two trends of the (mostly failed) goals of artificial intelligence:

  1. to build well-bounded microworlds in which every variable is carefully controlled and accounted for;

  2. to build expert systems that act in limited ways in the real world.

Edwards reveals the different philosophies of the two approaches:

Where the microworlds approach tried to build up to reality by gradually expanding a simulated world within the machine, expert systems workers carved reality up into a series of tiny "domains," capturing each one separately and worrying about recombining them later. Expert systems differed from microworlds in their focus on real-world problems, but they still reflected the closed-world belief in the reducibility of knowledge to a machine-like automatic system.

Contents Archive Sponsors Studies Contact