[JulianEdgar]

I wrote this back in 2008.

Long ago, even before I was a geography teacher, I studied how to teach it. The head of the geography department at college was a very smart person, and a brilliant teacher.

One day we were talking about teaching analogies and models, and the difficulty in simplification of knowledge without introducing straight-out erroneous ideas.

His example of the latter was: Clouds bumping into each other make thunder.

Much better, he pointed out, to say even to the youngest child: Thunder happens because of lightning.

In fact, clouds are a good example of these ideas. My little boy, who is 4 years old, asks what clouds are made of.

Tiny, tiny water droplets, I say.

So, how does rain happen, he asks?

I say: The tiny droplets run into each other and join together. When they are big enough, they fall to the ground.

While I am saying this, sometimes I think of a much more sophisticated model: water vapour, latent heat of evaporation and condensation, relative humidity, dew-point, hygroscopic nuclei – and other concepts.

A meteorologist would probably think of vapour pressure, a chemist might think at a molecular level, a physicist might consider terminal velocities, a climatologist might consider climate change, a minister of religion might think of God, an agnostic might think of the magnificence of nature.

In the description of clouds and rainfall that I say to my son, I am conscious of the gross simplifications I am making.

But that’s OK: every single thing I know about the world is a gross simplification of reality.

The intellectual models I use to make sense of what occurs around me are just reducible approximations of what really happens.

When I write technical articles, I am conscious that all the time I am presenting fundamentally simplistic models. I hope that they’re not of the ‘clouds bumping into each other make thunder’ type: but they may be.

Recently, I wrote an article on suspension roll centres, virtual pivot points and other ways of analysing suspension designs. In doing so, I consulted five different suspension design textbooks, and also considered very carefully the experience I have in developing human-powered vehicle suspensions, and modifying car suspensions.

As always, I was quite conscious during the writing of the article that the model I was presenting of reality was likely to be flawed: as I have already implied, every model we have of reality is, to a greater or lesser degree, flawed. However, I hoped that the information would benefit people’s understandings, especially in practical outcomes.

The day after finishing the article, I looked through a complex SAE paper on suspension roll centres. This paper immediately debunked several suspension ‘myths’, most of which I had implicitly or explicitly promulgated in the article I had written.

However, the paper was working at a level analogous to the ‘vapour pressure and hygroscopic nuclei’ theory of why rain falls: if I based my article on the SAE paper in question, perhaps less than half of one percent of my readers would understand anything I wrote. (If in fact I could understand it myself!)

So I could easily decide not to write anything at all: if it’s not ‘right’ and ‘correct’, surely it shouldn’t be written?

But that would be like saying to my son: I cannot tell you why rain falls; it’s too hard to understand.

I cannot tell you what a roll centre is; it’s too hard to understand.

Or I cannot tell you what a voltage is; it’s too hard to understand.

I cannot tell you what engine detonation is; it’s too hard to understand.

I cannot tell you how a tyre behaves when cornering, it’s too hard to understand.

And so on.

And these things – and all things – really are too hard to understand… if you want as ‘correct’ an understanding as it is currently possible to have.

Are my articles full of errors? So by definition, very likely.

Anyone who suggests that the technical articles they present for general readers are perfectly correct – or do not mislead in the slightest – just do not understand the nature of knowledge – and how all our descriptions of what goes on around us are just relatively simplistic models.

Me? I try to use the simplest model that’s consistent with not being grossly misleading…

(And 'using the simplest model that’s consistent with not being grossly misleading' doesn't include using a template to:

- Show where there is separated and attached flow on existing cars
- Guide the shape of rear extensions
- Show how rear spoilers on sedans should be positioned and shaped
- Allow the assessment of the ‘aerodynamic purity’ of cars)