Smart tools, dumb craft

Impressions

I have a new microwave, or rather I have an old microwave that is new to me. I don’t like it. It is bigger and more powerful than my old microwave. I didn’t need a bigger and more powerful microwave, but I don’t object to the size or the power. Actually I wasn’t entirely convinced I needed a microwave at all; I use it for very few things. Mainly I defrost meat, because I am good at putting together dinner on the fly but bad about planning ahead; and soften butter, because my daughter likes to bake, but, well, ditto.

I don’t like this microwave because it has too many single-use buttons and multi-step programmed procedures and not enough basic flexible options. I don’t mean that it is too complicated as such, because some tools have a lot of functions and thus need complicated means of interaction. The problem is that its interface is far more complicated than it needs to be, and sufficiently complicated that its design actually interferes with intelligent use.

Here’s an example. If I wanted, with my old microwave, to soften a stick of butter to cool room temperature, I could simply “defrost” it for about 20 seconds. I had only to lower the power and set a time. On the new one, there isn’t a simple way to lower the power; there are only options for various specific foods and purposes. So I have to press “soften/melt,” then watch a scrolling digital readout asking me to press a number for whether I’m softening or melting, and then another number for what sort of food item I have, and then a third for how many sticks of butter. The old process required me to press four easily readable buttons (defrost, 2, 0, start) and worked perfectly, because I’d experimented a bit to see how long it took to soften a stick of butter. The new process takes a good ten seconds to get started and halfway melts the butter, so that I have to stand and watch it through the (typically streaked and greasy) glass. And if I want to soften half a stick, I’m out of luck. Same for defrosting less than a pound of meat. It simply isn’t an option. And while there may be some way to make the machine do what I want, I’ll have to find a manual somewhere online to figure out how, because the “custom” settings aren’t.

I was thinking, yesterday, about how I would solve this problem. One way would be to plan ahead and/or just use the gas stove, but that’s not really in keeping with the spirit of the age, so take it as a design problem. Here are some observations:

The devil of false precision

Impressions

Eating lunch today I noticed on my bottle of soy sauce the words expiration date on label and, an inch away, a dot matrix stamp: 2019.03.28 14:48.

I expect that the stuff was bottled on March 28, 2016 at 2:48 pm and that it’s supposed to be good for three years from the date of bottling. But that’s not the same thing as saying it’s good until March 28, 2019 at 2:48 pm. Certainly a machine can record the exact time of bottling, but the idea that the soy sauce is good for exactly three years, for three years down to the minute, is absurd — as if, at twelve minutes to three on a particular March afternoon two years from now, the contents of the bottle will instantly develop a fuzzy blue mold and smell distinctly of gasoline. Obviously that’s absurd.

For one thing, it was bottled in Taiwan, so it would actually expire at 1:48 am EST and not in the middle of the afternoon.

“Three years from date of bottling” means three years, give or take. Give or take what? That’s the question. Six months, maybe? I would assume that they kept a bottle around for three years and it seemed to be okay. I doubt it’s very scientific at all. But it’s so easy just to take the present time, add three years, and stamp it on the bottle.

The expiration date on my soy sauce is not in itself a big deal. (I’ll use it within a few months anyway.) But this kind of arbitrary precision is everywhere — the practice of assigning a number to something, giving it as many decimal places as we can, and then slapping it on a label, noting it in a chart, entering it into a database — where it takes on a kind of magical invincibility, a rightness that can no longer be questioned or challenged. There are cases where this might have disastrous consequences, but more important is the impression of invincibility. Knowledge is power; false precision is an implication of knowledge; therefore false precision is an assumption of power. False precision is one way that science and industry and government claim power over us. But wallpapering the world with false precision builds false confidence in our own abilities, individually and collectively.

Every measurement is an estimate. If I were king of the world, I’d decree that every published measurement must be accompanied by a margin of error, e.g. “Expires on 2019.03.28 14:48 ± 6 mos.” It would be honest, it would be accurate, and it would remind everyone many times a day of the limits of human knowledge.

(And no, since you ask, I cannot think of anything better for a king to do than to demand accountability and humility from the powers of the world. Can you?)