In this podcast I discuss assumptions and how they are often hidden and can often mess up your day and your design. Assumptions are everywhere and usually they are not mentioned. The worse case of the problem is if these are hidden in a software algorithm and you are not aware of when they may be important. That is why I try to talk principles on this blog.

I use several examples. The first example is the previous post about buoyancy of pvc pipe. I did not include the end caps in the calculations. I knew it would change the calculations, but on a 10 ft piece of pipe it would be minor. Besides, the calculation of this would make an already complicated subject more complicated. Both the weight and the displacement of the end caps would have to be included and both may be difficult to obtain. The assumption was known to me, but not mentioned because doing so would have added another boring factor in an already lengthy post.

I also talk about how assumptions can be hidden in both the formula used and in the assumptions about the physical facts. This is especially important if you are extending the data far past the known points, extrapolation. The example I talk in length about is a bullet exiting from a gun. The path of the bullet is actually a curve, but it is “modelled” by a line in the gun sights. It also is affected by the weight and shape of the bullet as well as the charge used in the cartridge. An excellent written discussion of this can be found at Trajectory and Sight-in Relationships.

Another example of assumptions are interest rate curves on investments. I show an example of those in the picture at the top of this page. Because these curves are not lines but a special sloping relationship called the exponential curve small changes in the interest rate assumption can have a very large effect over time.

Finally, I also describe how to test for assumptions when using design software. Again, this requires that you have a basic idea of what kind of numbers to expect on simple cases before you try complicated real-life problems.

Gary

[…] the learning curve. In the beginning it looks a lot like the interest rate curves I showed in Episode 55. Learning at this stage is tough and lots of work to gain little because you have yet to have […]