Information Technology Reference
In-Depth Information
Complexity of Systems
We must have an understanding of the concept of complexity in systems.
It is safe to assume that the construction and possibly maintenance costs
of complex systems are higher than simpler ones. Before embarking on
the building of any system, one must be able to judge to some extent
whether one is dealing with a complex system or a simple one.
Organizations are complex systems based on a combination of mech-
anistic, as well as social systems that are organic and self-organizing.
Because of this plurality, one must apply a combination of thinking
approaches. Mechanistic subsystems, such as manufacturing or distribu-
tion, benefit from linear thinking, analysis, and the dissection of processes
into smaller steps. Systems thinking helps one understand the dynamic
nature of relationships. Thus, to assess the complexity of any organiza-
tional system, one must not only look to the outside environment for
observable external behaviors, but also understand the basic assumptions,
beliefs, and value systems that sustain the organization, its structures, and
these behaviors.
The pertinent question that now arises is whether there is any rule of
thumb to follow in an effort to determine if a system one is trying to
create is simple or complex. One rule is that simple systems are based
on simple processes. Of course, as with any rule of thumb, there will be
exceptions. If there are a number of interacting subcomponents, one
cannot consider a system simple. A possible exception occurs when the
subcomponents are homogeneous and are governed by uniform laws. For
example, we associate properties (density, pressure, mass) to various kinds
of matter (solid, liquid, gas) without having to deal with constituent
particles at the atomic or subatomic level.
Simplicity is sometimes associated with determinism. It is natural to
expect that nondeterministic systems, or those systems that cannot be
modeled by standard mathematical methods, are complex. But the con-
verse may not be true. The famed Rubik's Cube is deterministic, but not
simple by any stretch of imagination.
Another way to classify complexity, and one that software professionals
are familiar with, is on the basis of whether the system is centralized or
distributed. Complexity experts maintain that distributed systems can do
all the things that centralized systems can do. In fact, they feel that the
centralized system has some pitfalls:
They need a central intelligence to manage them.
The central intelligence is susceptible to infiltration and interjection.
It may be difficult to get accurate and timely information to a
central point.
