Systems
The Next Person Who Touches It
I spent the '90s fixing Mercedes door locks that someone else broke. Every design decides who pays when the next person touches it, and the good ones admit the trade.
I spent a chunk of the ’90s fixing Mercedes door locks. I was paying my way through school as an auto electrician at a Mercedes dealership, after a few years installing car stereos and alarms.
Older Mercedes door locks were vacuum-powered. You didn’t flip a switch. You lifted or pushed down the lock, and a few seconds later the other doors followed. A lot of owners wanted remote locks, usually at the same time as an aftermarket alarm. So an installer would try to drive the vacuum pump straight through its control wires, without understanding the logic of how the pump worked, and often broke it. A few weeks later the car would be in my bay.
The installers weren’t stupid. I’d been one. They were the next person who touched it, and nothing about the system told them how it was meant to work.
That’s the whole problem, and I think it comes down to one rule. Every design decides who pays when the next person touches it. The good ones spend in proportion to what a failure costs whoever’s next.
The cheap rung
My favorites to work on were GMC and Chevy trucks and Hondas, because they were easy. A bad connector pin meant cutting off the whole plug and splicing on a new one. That sounds crude, but a failure on those trucks cost a splice and an hour of a tech’s time. Cheap failure, cheap design. GM made the right call.
The fixable rung
The cars I thought were the coolest were the Mercedes, and they were not easy. On a Mercedes you could take a connector apart and replace a single pin. Who does that? Someone at Mercedes had designed the electrical system for me, the technician they’d never meet, years after the car left the showroom. On an expensive car built to last fifteen years, that paid for itself.
They also knew exactly whose life each choice made easier. To get inside the driver’s side of the dash, you first had to take down a circular corrugated steel partition. I hated it. Then I figured out why it was there: in a crash, it kept the driver’s legs from breaking against the dash. Mercedes made my job harder on purpose, because on that decision I wasn’t the person they were designing for. The driver was.
The rung where the mistake can’t happen
The next place I saw engineering at that level was NASA, on Space Station assembly.
The EVA connectors had bail handles, cam-driven latches, and heavy alignment keys, so it was hard to plug one into the wrong place and easy to mate it with suit gloves on. Now imagine bending a pin on a big multi-pin connector halfway through a spacewalk, wearing those giant gloves. How would you even fix that?
You wouldn’t have to. Those connectors were designed so bent pins just didn’t happen. In my time there, I don’t remember a single bent pin on a station spacewalk connector. NASA’s own records say where bent pins do come from: the ground, from people mating connectors wrong. That’s my alarm installers, designed out before anyone could make the mistake.
Those connectors were not small and not cheap. But neither is another trip to the space station to fix a broken part. In orbit, the expensive connector was the cheap option.
The trade nobody admits
So GM wasn’t wrong and Mercedes wasn’t simply right. I make the GM trade myself all the time. Shipping now versus building it right, cost versus good enough: that’s the job. Anyone who says they always build like Mercedes either has an unlimited budget or isn’t shipping.
Every engineering discipline beats the same lesson into you: it’s always a trade. What are you optimizing for, and what are you giving up to get it? You can always make a better design. The real skill is knowing when enough is enough, and for whom.
In software, the dash is a codebase. Most bad engineering I’ve seen wasn’t the cheap trade. It was the cheap trade made without anyone admitting it was a trade, so the next person found out with the dash already torn apart.
So the lesson I took from that dealership isn’t “always overbuild.” It’s three questions I still ask:
- Who touches this after me?
- What trade am I making, and am I saying it out loud?
- Who pays for it later, and do they know?
That alarm installer never got told. The trade was made years before he picked up a wire, and nobody wrote it down where he’d see it. He found out the way the next person always does: by breaking it, and then it landed in my bay.
I learned to ask those questions on my back under a dashboard. I’m still not sure I always answer them honestly. When you cut a corner, do you tell the next person?