The Policy Forum article "The Y2K Problem" by Robert F. Bennett (Science's
Compass, 16 April, p. 438), was an excellent follow up to the special issue
of Science on Complex Systems (2 April). Within the next 12 months we
shall learn just how interconnected and interdependent the institutions of
contemporary societies have become.

Consider the fact that a major manufacturing firm such as General
Motors has thousands of suppliers. Each of the companies supplying parts
or services to GM also has suppliers. And those suppliers in turn have
suppliers. Suppose that GM is completely year 2000 compliant and the
majority of its suppliers are as well. But, for whatever reason, GM
cannot obtain 5% of the parts needed to build automobiles. What will GM
do? They will have to stop building cars. And all of the compliant
companies shipping parts to GM will lose an important customer.

In terms of complexity theory an interesting question is, What
fraction of the companies in an interconnected society need to fail in
order to cause the failure of nearly all of the organizations in the
society? The Gartner Group has estimated that in many countries around
the world 50% or more of the companies in the country will have at least
one mission-critical system failure. These countries include Russia,
China, India, Indonesia, Japan, Germany, Turkey, Saudi Arabia, South
Africa, Venezuela, and Argentina (gartner5.gartnerweb.com/public/
static/aboutgg/pressrel/testimony1098.html). That is a lot of countries,
a lot of people, and a lot of supplies shipped to the United States.

Also, according to the Gartner Group the countries most prepared
for dealing with the year-2000 computer crisis include the United States,
the United Kingdom, Canada, Australia, and the Netherlands. In these
countries 15% of companies are expected to have at least one
mission-critical system failure. Which societies will be least affected
or will recover most quickly - less developed societies with less
dependence on computers and automated equipment or advanced societies that
have been working hard to repair date-sensitive equipment?

Another "complexity" issue involves the many chemical plants,
refineries, nuclear reactors, and pipelines we have constructed. All
facilities that handle hazardous materials are designed with back up
systems. How many programmed logic controllers within these facilities
need to fail in order to cause a spill? Will some societies shut down
their dangerous equipment and others not?

It seems that we are about to witness an "autopsy" of modern
society. As one system or company fails, we shall learn what other
systems, companies, and countries depend on it. The year ahead will
present unique opportunities for education, research, and public service,
as we help the public understand what we are experiencing.

Stuart A. Umpleby
Director, Research Program in Social and Organizational Learning
The George Washington University
Washington, DC 20052, USA
E-mail: umpleby@gwu.edu