I heard someone say that to me recently as they entered another weekly test cycle and I pondered whether testing could be measured in Tonnes (or mass). And since mass and energy are irrevocably connected, I wondered whether it could be true as we are happy to talk about the energy expended in testing. The SI definition of a Kilogram changed late last year from the physical platinum prototype kept under lock and key near Paris to a calculated value based on Planck’s constant. With this new definition of the Kilo a precise definition of Planck’s constant is required, that is 6.626070150×10−34 Joule-seconds. So does it help to plagiarise his equation for our purposes?
Planck’s constant (h) defines the value of the relationship between frequency (f) & energy (E) in particle physics, specifically light.
E = hf.
How does this relate to application testing?
Well the energy expended in testing is certainly related to the frequency (f) of test cycles. The higher the frequency the great the energy. But similarly, we are not typically talking about testing one particle, it is a compound structure with a greater mass and hence the Energy is also greater.
Clearly with software testing Planck’s minuscule value is in appropriate so we need a new value and we need a standard form of energy. As we know, most testers run on alcohol. A glass of wine contains about 100 Kcal and weighs 125 grams. The average person burns about 100 Kcal per hour working at a desk job, so handily they need one glass of plonk per hour to test.
Now, this is the bit I don’t know. What have you observed? How long does it take to test a particle, or a single unit in other words? Is it reasonable to say an hour as an approximation at this stage?
Hence our ‘Plonk’ constant based on the measurements is (staying in hours as SI not very helpful on this scale).
h = E/f = 0.125 kg/1 per hour = 0.125 Kg-hours
So a Tonne of testing requires 8000 hours and is physically impossible in a week for one person, after all it would take about sleepless year. But if it was an automated regression test, that could easily have touched 8000 units and hence that could be realistic. Equally, while it is running you don’t need any bodily fuel but you might chose to put your feet up and have a glass – does that count?
So I think anyone who has a “Tonne of testing to do this week” either needs a much bigger team or some technology to help.