I wonder how beneficial switching to the metric system would really be. It's trivial to convert between systems if need be and pretty much all serious engineering/science work (where confusion can be costly) is done using the metric system these days anyways.
(Numbers made up, based on what I read many years ago on [1], but most of the links are now dead.)
Situation: mother and 6 year old boy in hospital in the UK. In the UK, everything is metric except casual discussion of body weight and height, and road distances and speeds. And beer in pubs.
Doctor: I'm prescribing this drug, the nurse will give it to your son.
Nurse to boy: can you stand on the scale please? OK, 45… that makes the dose 5mL.
Did you notice the error? A 6 year old won't weigh 45kg, the digital scale is somehow set to pounds. The boy actually weighs 20kg — he's about to receive a double-dose.
Neither the nurse nor the mother (or the boy for that matter) noticed the incorrect weight, and the boy was seriously injured (perhaps killed, I forget) through an overdose of the drug. The NHS then replaced any scale that could give a reading in pounds, to prevent a repeat of the error.
In the rest of Europe, nurse or mother would have known 45kg wasn't a reasonable weight for a 6 year old boy, and investigated the problem.
I'm sure you can think of other technical or semi-technical situations where unfamiliarity with the metric unit can lead to mistakes.
This is particularly surprising given that most people in the UK use “stones” as a measurement for weight, so pounds would be just as confusing as kilograms.
The imperial system has shortcomings for small units, which means that regular people will encounter grams/millimeters/milliliters in their daily lives, so they are effectively using a hybrid system already, which is just the worst of both worlds.
Yes, converting is technically trivial, but you still have to reach for a calculator to do it.
I wish the metric system didn't have shortcomings in the medium units, though. There's no (commonly used) parallel to feet in metric. Centimeters are too small for the kind of estimations feet are useful for. That's really the only imperial unit I care about.
2 feet is about 600mm or 60 cm or 6 decimeters or 0.6m or 0.06 dekameters. Just pick whatever prefix that gives you the number of significant digits you want.
Metric doesn't "lack units", because there's only a single unit for every measure. Instead there's prefixes for every power of ten around human-scale measurements, and for every power of thousand for much smaller and much larger measurements.
You can use decimetres if you wish (0.1m or 10cm, so about ⅓ foot). In practise this isn't really a shortcoming, so hardly anyone does use them.
About the only place I see decimetre markings is depth gauges on rivers/canals/bridges, and rulers used by geologists when photographing a rock formation.
>The imperial system has shortcomings for small units
Don't forget the medium and large units!
I guess miles on their own work fine, but it'd be nice if someone could say "2000 feet" and I didn't have to think about how many miles that means.
And our two speed measures of feet per second and miles per hour are close to equivalent (1 fps is about 0.7 mph) but nobody knows that to convert it in their heads. It just happens that 3600 seconds is the same order of magnitude as 5280 feet.
Not that metric is totally better in that regard since they're stuck with minutes and seconds too. There's 86,400 seconds in a day, maybe when we're making this big calendar change we can switch it over to 100,000 seconds at the same time.
Or we could change miles to be 3600 feet so that feet per seconds and miles per hour are equivalent. Added bonus, that's not too far off a kilometer.
0.7 and 1 are not close to equivalent even when speaking informally. The difference is the same as driving normal highway speed vs license suspension (or jail) speed
One thing I learned from a Disney movie of Goofy driving is that multiplying a speed in miles per hour by 1.5 gives a surprisingly close approximation to feet per second.
Just one example: when I’m working on a random unknown bolt, I have to try two sets of sockets on it because I don’t know if it’s metric or imperial. Worse, there are some sizes which are close enough to grip and seem like a fit, but which will damage the head when more than moderate torque is applied.
There are lots of little things like this, multiplied by millions of people.
The imperial system is a problem in daily life. Most of my American friends don’t know how many ounces are in a pound or how many teaspoons in a tablespoon. This makes price comparison at shopping very difficult and cooking is also harder. Sure you can survive but imperial is just an incredibly inefficient system.
it really doesn't help that you being non-american, could conceivably be talking about any one of about 20 different 'pounds'. America itself only uses the avoirdupois pound, but if you thrown internationalism into the mix, it could be a troy pound, an ISP or a non-US avoirdupois pound, maybe we're even talking about a russian pound... and that's all before we get to pound-mass vs pound-weight/pound-force
here, let wikipedia muddy the waters even more for us:
Comparison shopping is difficult because the stores are allowed to get away with marking interchangeable items with wildly dissimilar units (type and/or magnitude), quoting per item, per ounce, per pound, per gram, per quart, etc., all on items that are sitting next to one another. All of the grocery chains that I've ever shopped at are guilty of this.
People who are serious about cooking know about the volumetric units used in cooking and the relationships between them. (teaspoons/3 = tablespoons)/16 = cups.
But your American friends who don't know how many ounces are in a pound? I hope you're kidding. That's basic stuff, taught in elementary (primary) school.
I am not kidding. Most seem to be resigned to the fact that it’s 12 or 16 and that’s close enough. Also nobody knows how many cups are in a pack of flour so buying something for a recipe is hard too.
These comparisons are much easier in the metric system : it is either per unit or per kg.
If it is per gram, it's easy still because *1000 or /1000 is very easy. (basically you just move the decimal point, no calculator needed)
It is because we count with base 10 numbers that metric is easy and imperial is difficult.
Another related thing is the nutritional value info box that every food item in the US is required to have. That's a good thing in itself, except the US allows this info box to show the values "per serving", which is usually a completely arbitrary measurement, making it impossible to compare two food items to determine which has more sugar, or which has more protein.
In the EU, the same thing exists, but it is always per 100g or per 100ml. No exceptions. So comparing two food items is always super easy. I don't know why the US allows this idiotic loophole of per serving.
The fact that converting measurements is trivial in theory doesn't mean that you don't have those measurements embedded in physical artifacts in non-trivial ways.
Any sort of machinery that deals with measurements, has some sort of "standard" embedded in it.
I'm a woodworker, so I'm going to pick on the world of woodworking.
A thickness planer (thicknesser if you're from the UK) dimensions wood to a consistent thickness. The adjustment is typically done by raising or lowering the bed relative to the cutterhead. There is generally a crank or handwheel that turns some kind of gear or screw mechanism that in turn moves the bed.
In the US, woodworkers generally use inches, and work in thicknesses that are some even number of 16ths of an inch.
It would be possible, in theory, to set any planer for any dimension in the continuous range of dimensions it supports. In practice, the adjustment mechanism is set so that a whole turn of the crank corresponds to some whole number of 16ths. Moreover, this is usually set up so that a whole number of 16ths lands at an easily repeated position: crank handle at 12 o'clock or 6 o'clock. This makes it blindingly easy to repeatably hit the same mark (within woodworking tolerances) every time you use the machine.
So far, so good.
The shop I belong to where I have access to a planer has a modern Powermatic planer. Powermatic no longer manufactures in the US, and the planer was a bit of a source of mystery to me until I dug up the docs and read them. One full turn of the crank is 1.5mm. I'll spare you converting: That's 1/16th of an inch, less 3.5 thousandths of an inch (thou, rhymes with cow).
That doesn't sound like much, but it adds up. Over 16 turns of the crank, you're now 56 thou off where you expected to be, or, almost exactly 1/17th of an inch[0]. Moreover, hitting 3/4" exactly (within woodworking tolerances) requires a bit of guesstimating about how much extra you need to turn the wheel past 12 o'clock or 6 o'clock.
Non-solutions to this problem:
1. Put a measuring device on the machine. They either aren't accurate enough, don't stay accurate, or cost a fortune (i.e. anything digital that is both accurate and stays accurate)
2. Work in metric. I would, happily, except for the fact that literally everything else in the shop is inches. Chisels in the US are an even number of 16ths of an inch wide. A dado stack for a table saw cuts dadoes an even number of 8ths, 16ths, or 32nds of an inch wide. The simple act of finding a measuring tape that's metric is a pain in the ass. I can go to Home Depot, Lowes, or any hardware store and find a dozen options in inches. If I'm lucky, there's ONE with metric, and it has inches on one side of the tape.
But wait, it gets worse if you're a metalworker. Metal lathes have a leadscrew that rotates a fixed number of rotations per rotation of the workpiece (non-continuously variable by gearbox). Those are threaded either in either metric or inches. To do metric work on a lathe that's natively inches (or vice versa), you need gears in the ratio of 254:100, reduced to 127:50. If that sounds like a lot of teeth on one gear, that's because it is; making that set of gears requires either an impractically large gear (won't fit on the lathe), or impractically small teeth (incapable of transmitting the required torque)[1].
Once you start making actual physical things, you find that your system of measurement is embedded in nearly everything around you in ways that are difficult to work around while maintaining sufficient accuracy and precision.
[0] Yes, this means that 1.5mm is almost, but not quite, exactly 1/17th of an inch.
[1] There exists a very close approximate solution that can be used subject to limitations you'd need to find a machinist to expand on.
When I did a little home improvement project I always thought what a PITA it is to multiply 5 and 3/8 inches seven times. Let alone divide that number by 3. Metric is so much easier.
Serious science experiments like the resoundingly successful controlled descent into martian terrain.
IMHO this is just the tip of the iceberg. I can’t even begin to fathom the inordinate amount of conversion mistakes there might be but we never hear about or even detect because this is not NASA yet have possibly damaging consequences.
The lander crash could have just as easily been happened using 2 different metric units. It wasn't caused by improper conversion, it was caused by no conversion. Each part assumed the other was using the same unit.
> The lander crash could have just as easily been happened using 2 different metric units
Nobody uses anything but meters and seconds with metric in science. But even if - using 2 different metric units would give answers differing by a factor of 1000 - a little easier to notice than Pound vs Newton (1 to ~4.45).
> Nobody uses anything but meters and seconds with metric in science.
A lander is engineering and science, and I can guarantee you that there is a non-zero probability of one part of a system working in m/s and another in km/h — which would lead to the exact same error.
>Despite this being a "solved problem"(TM), it's not really trivial in any meaningful sense, neither for regular people nor for developers creating a system that should deal with units in both systems.
It really is trivial for "regular" people as everyone has a device in their pocket that can quickly convert to any unit you want (which is an infrequent thing to need in daily life). As for developers, there's lots of changes we could impose on the world to make our (developers) lives easier but easing the burden on developers should rarely be a consideration when considering such changes IMO.
Well, by that metric it's about equally trivial for any maths student to solve indefinite integrals (e.g. using Wolfram Alpha), right?
It's not that I disagree about the utility of modern technology, but I just think that some things are inherently difficult for humans even if there is a handy oracle to compute them.
