I think, much like the “elevator” test interview question, it’s too hard to not get tied up in domain knowledge problems. What I find interesting is a point I made yesterday to my help-meet about the test kit. I called it a “machine”. She said it’s not a machine, which is perhaps valid because it has no moving parts. But your wrist-watch (of any type) is a computer, it just may or may not have a lot of moving parts, or it may have a microcontroller or it may have a general purpose processor. Every wrist-watch is still a computer. The covid test is a precise instrument, just mass produced with the aim of protecting it’s function against the user to achieve some level of accuracy. So testing that accuracy will always be out of scope for us.
The home test kits are machines, they just use chemical action. and so it’s no surprise that the use instructions are very careful around all mechanical steps that reduce chemical contamination or interference. As an engineer, I find the instructions the most fascinating and the most testable part of the entire thing. Intended instructions are written so as to make the results uniformly reliable. The kits available here look identical, but work differently, so that was interesting too.
- I kind of have to drop some domain knowledge in here though, the lateral flow test (the ones distributed in the UK uses a swap technique , which is a user barrier, but improves early detection.) is not the only “simple” test. Another test kit called LAMP is also possible, but no marketed. It uses the same amplification technology as the PCR (Polymerase Chain Reaction) lab tests, but, and here is the shocker, it takes half the time of a PCR, and you can technically do it on a stove-top with a thermometer and a pan. I’m sharing this little point, because, whenever we test anything, we have to use comparisons to other similar systems in how we judge not just the visible quality, but also efficiency.