A young bull lived in a small herd in a nature reserve on the shores of a lake.

He felt strong, but was, like any young animal, at the bottom of the hierarchy.
One day he challenged a senior bull, it was a mistake, he was expelled from the herd.

The nature reserve was small, so he could not really leave, but returning to the herd was impossible too.
He took the gamble and swam into the lake, quite far away there was a small island. There he was safe from the herd.

A young student was working on an ecological computer model. Can you predict vegetation in nature reserves near the water even if the water becomes saltier and will flood the nature area more often?
A large amount of data was available, of vegetation present at thousands of locations, of many years.
A small island gave a big anomaly. The vegetation was completely different than you might expect.
The one bull had changed the ecosystem of the small island completely.

Is vegetation growth predictable with a computer model?
Roughly speaking it is.
And a vastly improved computer model, would it be able to predict the details of the plants growing in all areas?
Not a chance.
To predict the vegetation of this island, the model shoud have contained the entire workings of the psyche of the calf and that of all his herd members.

A computer model does not have a chance, the whole of science does not have it, to make predictions in such a detail.

I finished my job in this research and started the next:
What are we able to predict and what not? Where are the boundaries of what we are able to know?
Where does our reason cease to be useful?

The smallest, first step is to recognize how much we do not know and try to map it.
To start with that, we could make the not-knowing visible.
When we get used in doing that, I believe we will see a pattern.
Until finally we can create a map, a mental map with what we know and what we do not know.
And we can make it into a science. Just like the rest of science, we will not get those boundaries absolute, but we will change it, by experience and theory.

It will help us to give us an idea where we can book progress, with science or computers and where it is unlikely.
It will also make it easy for us to visualize in science and computers where knowing ends and not-knowing starts.

In fact, I see it as a task for software developers and scientists, to show the boundary between knowing and not-knowing in their software or scientific field, so it can not be used for tasks where they are incapable of.

How you could do that, I'll tell you in the following blogs. Subscribe to get them in your mailbox.

Also read blog 1 in this series: What the hell is pu?