Re: Defining a system

["Vic Forte" <vic@vicf.com>]

> I would suggest we accept the previous as the definition of a system, and
> discuss clarifications of what this means and implies.

OK. Let's have a go.

> Some questions I have are:
> 1) Can a component belong to more than one system?

Yes. A system is no more than a way of framing reality. There is no such
"thing" as a system "out there" - systems thinking is a way of looking. You
can frame reality in different ways and components may belong to different
systems depending upon how you frame them. For instance the component Wayne
is a component of the system "Wayne's employment" - and also part of the
system "DEN".

> 2) Does a system have inputs and outputs?

Yes. Open systems (systems which interact with their environments) have
inputs and outputs. They commonly have other features relating to this, such
as purturbation, or noise and feedback (see below).

> 3) Does a system ignore its inputs and outputs?

Usually a system takes its inputs and transforms them into outputs. This is
a simple system. A higher order system has feedback and is able to take an
aim as one of its inputs and by comparing outputs achieve homeostasis
through the adjustment of one or more inputs against the aim. This is known
as negative feedback.

Higher orders of system do not only recieve an aim from outside but are able
to evalute the aim and also determine their own aim for themselves. This is
what Dr. Deming meant by "the aim is a value judgement". In higher order
systems aims are determined by values.

Even higher orders of system can try to predict the future, and also
redesign their own internal structures. This was the kind of system that Dr.
Deming talked about in his famous systems diagram.

> 4) Does every component listed in the aim belong to the system?

The aim gives identity to the system. The effectiveness of the system in
meeting its aim is a function of the congruence of the systems elements.
Complex systems require complex forms of co-ordination and control in order
for their components to achieve congruence. The relationship of components
to the aim of the system may be very clear and simple but also may be subtle
and indirect.

> 5) Does every component of the system need to be listed in the aim?
>

The system can have one aim for itself, and components can have subsidiary
aims, differently defined which may contribute differently towards the
primary aim. These subsidiary aims would not be listed in the primary aim.

The aim does not reside equally in every part of the system, nor does every
component of the system contribute equally to the aim. Aims can change. Aims
need to be communicated. Where the components of a system are intelligent
beings, aims need to be shared in such a way that each component part can
see where they fit in and how they contribute.

> I hope we can accept Dr. Deming's defintion for a system and only propose
any other definitions if there is something found lacking in the current
one.  That would allow us to concentrate on what the implications of a
system are.

Well. I think Deming taught "appreciation for a system". I do not believe he
claimed to have the final say when it came to systems thinking. What I have
said above draws from general systems theory and is not specifically
mentioned in Deming's work. Most of it is sufficiently fundamental, however,
for us to assume that he would have accepted these ideas.