It looks to me like accounting and mathematics have the opposite approach to explain things.

Math : defition -> property

In mathematics you start by defining the objets you work with. Then, out of the defintions you deduce interesting properties. For example : You define what a bounded sequence of number is. Then you can prove that you can find a subsequence out of it that converges.

The advantage of this method is that there are no exceptions. The problem is that sometimes a property only holds for an undefinable subset of the objects matching a defition : “it almost always works”. In that case you must use approximation and heuristics.

Accounting : property -> definition

In accounting you start from the idea of a figure that represents some important aspect of reality. Then you determine a calculation method to obtain a number which approximates it. For example : you want to determine the value of all final goods and services produced by an economy, aka GDP. You can approximate this by the money spent by households on consumption, government expenditure in infrastructure, corporate investment in production means … To remove the money spent on goods produced outside the economy (ex: prime materials, machinery…) you remove the sum of all imports.

Although this is a logical reasoning, contrary to mathematics you cannot state a fact out of a formula. Even if GDP has declined it may not mean the economy is less productive. In this sense accounting looks a lot more like software engineering. If on your application monitoring dashboard you see that cpu and memory consumption increased with the new release, it might not mean the code is less efficient.

Implications