Uneasy Modelling

‘Essentially all models are wrong, but some are useful’ – George Box

I heard this quote during a webinar on modelling cycling physiology and its relationship with power output. While the quote is related to statistical modelling, it immediately made me think about the process of developing programming object models.

Sometimes modelling is hard. I often sit with an uneasy feeling that the model I have just isn’t right. Other times, modelling is easy. The right model just seems to fall out. Most times, it’s somewhere in between the two: an initial model is developed with areas that give rise to that feeling of unease. However, with a little patience and some test-driving, the creases can be ironed out and a neat solution emerges.

To help with that uneasy feeling, I now fall-back on this quote and remind myself that a model doesn’t need to be perfect to be useful and reliable. Furthermore, it reminds me that when I might rest on my laurels, complacently thinking that a model is perfect, there might be something that I’ve missed or that can be improved.