Copyright: 1975
Publisher: John Wiley & Sons
ISBN: 0-471-92563-2
This book is a bit of a classic and in fact I found out recently that it was reprinted again in 2001. I read the 1975 version that the local university had in their library and while the book is over 30 years old, it still has a tremendous amount of value with regard to the thought processes that go into information systems. What follows is going to be less of a book review and more of a collection of random notes that I took as I read the book.
Simplification
The Square Law of Computation says that the number of computations required to solve a system rises at least as fast as the square of the number of equations in the system. The way that Newton was able to even come close to solving the problem of predicting the orbits of the planets was by simplification. Rather than trying to compute the gravitational forces of all the planets along with all the moons and asteroids present in the Solar System, he reduced the problem to just the planets and the Sun. He further reduced the complexity by ignoring the interaction between the planets and focused only on the interaction of each planet with the Sun. By doing this he made it possible to perform the calculations necessary to predict the oribts using just the mathematical tools he had.
The danger of over-simplification can be seen in the Echo satellite problem. The Echo was an inflated Mylar sphere that kept giving programmers fits as they tried to predict its orbit around the earth. All the classical formulae were giving inaccurate predictions and no one knew why. Finally it was "discovered" that the sphere had such a low density that the sun's light radiating on its surface was throwing the orbit off. What could normally be simplified away in other orbital calculations became significant due to the very low density of the Echo.
Max Planck on Scientific Truth
In Scientific Autobiography Planck stated - "A new scientific truth does not triumph by convincing its opponents and making them see the light, but rather because its opponents eventually die, and a new generation grows up that is familiar with it."
Reichenbach on making a fool of yourself
"Indeed, the willingness to make a fool of oneself should amost be a requirement for admission to the Society for General Systems Research, for this willingness is almost a prerequisite to rapid learning." Meaning you have to be willing to try a few ideas out without complete knowledge if you are going to discover truths... at least in a rapid manner.
If we want to learn anything, we musn't try to learn everything
In this world there are things called "stepladders" and things called books. If you are in a library with no ladder and you want a book on the top shelf, what do you do? If you step back and stop looking at books as "things to be read" and think of them as "things that can be stacked to give me more reach just like a ladder" the solution is obvious. If everything we learn is "hard and fast" and there is no ability to look at things from a different angle, we reduce our ability to solve problems.
Observation and Behavior
When anayzing a system, it is impossible to observe every state of the system. The idea of General Systems Thinking is that you reduce the complexity of the system to some functional notation that you can then extrapolate onto the system. This gives us the ability to predict the behavior of the system without having to observe every possible state. One problem is if you characterize the system based on one observed state.
If someone spends their life as a doctor or a cook, we tend to classify them as "physicians" or "chefs" because these are the behaviors they exhibit most of the time. However if a student were an honor roll student for 3 years straight but one day decided to set fire to the library, most likely they would still be kicked out of school. They are characterized by that one incident to the detriment of their educational career. This is a good thing in the case of the student but can be a bad thing when you are trying to analyze a system and happen to characterize the entire system based on one observed state of behavior. The classification of behaviors can be characterized as follows:
1) Typical behavior or Exceptional but important behavior
2) Average behavior
Which classification system you choose depends on the impact of the deviation from the norm. The test is in the impact of the single observed state. If a drug heals 99 people of their flu bug but kills one of them... you can't ignore the one. But if the drug heals 99 people and gives one a bad rash... you can probably get away with just telling people of that small risk.
The Goal of the Systems Thinker
Towards the end of the book, Weinberg explores what it is that a Systems Thinker is trying to do. Ultimately it comes down to this. A General Systems Thinker is trying to improve our thinking, not necessarily solve the riddle of the Sphynx. Here are some questions that he suggests we should always be asking:
1) Why do I see what I see?
2) Why do things stay the same?
3) Why do things change?