Bonus: a story about incorrectly built models, for extra-loyal readers who got this far.

Many years ago I was enthusiastically exploring the capabilities of modern computers (at that time I had a Pentium 166 MMX with 16 MB RAM), and I learned about overclocking - a way to make the processor run faster. In the same context I learned about the importance of proper cooling, which was necessary to keep an overclocked computer stable.

There was a program called "Rain", that promised to make computers run cooler by invoking a special "halt" instruction when the processor was idle. In plain English, it would let the CPU rest when there were no tasks - that's my understanding of it at that time. People were discussing it on the forums and some were saying they could feel the difference in the temperature of the air exhausted by the fan.

Although I did not overclock my CPU, I thought that it wouldn't hurt if I used the program anyway. Merely running it wasn't enough for me, I wanted to measure the effect.

To do so, I kept my hand behind the computer (where the power supply fan blows the air out) and sensed the temperature. Then I started the program and observed how the air got cooler. This was the essence of my experiment; although I can't say I felt any difference, I made myself believe there was one (after all, people on the forums confirmed this).

However, it was an illusion, because they were actually referring to the CPU fan, not the power supply fan. In other words, I was measuring temperature in the wrong place.

Now that I think about it, I doubt I would have felt the difference, even if I knew where to measure; a human hand is not that sensitive to small variations in temperature.

All of this happened before I was brave enough to disassemble my computer, I wasn't aware of the existence of another fan inside the big box. While watching others disassemble computers, I did not see other fans (this was the case with my previous computer with an Intel 80286 processor, running at 12 MHz, with 1.4 MB RAM); so I expected it to be the same.

In practical terms, I had an incorrect model of the system in my mind, and thus my experiment was flawed. This was a harmless error.

However, I wouldn't want to walk over a bridge designed by an engineer who had the wrong model of physics in their mind, or get operated on by a doctor with an incorrect model of the human body (think of Dr. Zoidberg).

Incorrect models are all over the place. They can be useful too, but only up to a point.

Pages: 1· 2