As a hardware engineer I've been primarily using metric for over three decades, and like the machinist quoted in the article have had to deal with bullshit of c

As a hardware engineer I've been primarily using metric for over three decades, and like the machinist quoted in the article have had to deal with bullshit of converting back and forth. It's not just about inches and centimeters. If you're working with small motors, the Ke (back-EMF constant) and Kt (torque constant) have different and crazy Reagan units (e.g. V/krpm, inch-ounces/A) but in SI they have units of V/(rd/s) and N-m/A, which if you crunch them down to kg, m, etc. are identical and so have the same numerical value (because they represent the same transformation from electric to mechanical or vice-versa). Power is the product of voltage and current or torque and speed - if you use the SI units no conversion required. Inertia was confusing as hell, some vendors specified it in MOISS, or milli-ounce-inches-seconds-squared, not only involving different unit but you can also get balled up in the mass vs force confusions. Converting a drawing from one system to the other perfectly is practically impossible. Conversions can't always be exact and because tolerances are commonly specified in round numbers within a system. In my garage I still have both metric and Reagan-unit tools. Even though my cars, even the American ones have been metric since the '80s, I have to have the other tools for other household things like the garage doors! And then there's the tire size abomination - an unholy union - a 255R70-14 is 255 mm wide at the bulge, the sidewall height is 70% of that, and the rim diameter is 14 inches. I agree that 1 °C is too big a step when dealing with thermostats but that's easily solved by using 0.5 °C resolution.