> Metric is already used in areas where it actually matters (e.g. STEM)
Using French Revolutionary units doesn’t really matter in STEM, either: one can conduct science just as well in any units one wishes. One unit of measure is not more scientific than another. For example, degrees Kelvin and Rankine measure the same thing with different units. If anything, the Rankine degrees are more precise!
You shouldn't use degrees for Kelvin, it's an absolute unit, the degrees are needed for the relative units like Celsius.
Anyway, the French system isn't what people mean by "metric" in this context, they mean the SI system of units, and so in practice it's not so much that it wouldn't matter which you choose as that you don't have any option except SI.
If you wanted an independent system of units you'd need to do a lot of expensive metrication, and in practice Americans are too cheap for that, so the US "customary" units are just aliases for so-and-so-much amount of some SI unit, they aren't actually independent at all.
The reason people focus on metric is that for everyday people that's the part which jumps out as more intuitive. All these nice powers of 10, very tidy.
> For T&E it really matters see NASA's Mars Climate Orbiter and the need for heroics in the Gimli glider.
> You need to keep to the same unit.
Completely agreed. You’ll get similar issues if you have one set of parts using m/s and others using km/hr.
And you’ll avoid those issues with any standard, whether it’s m/s, knots or mph. The important thing is to have a standard.
It's useful in STEM because SI units are often straightforwardly defined in terms of one another, it substantially reduces the amount of arbitrary constants you need in your equations, making them a lot easier to work with.