In Defense of Software Engineering
You’re a sailor on a ship. Fifteen years ago, the ship was a guy, a dream, and some logs lashed together, but since then, it’s grown beyond recognition. You joined when there were a thousand sailors aboard; now there are three thousand. The captain, of course, is still the captain. If he were ever replaced by one of those “professional captains” who’ve never built their own ship from scratch, you’d probably leave.
The ship is a monstrosity. It’s never had drydock maintenance, built entirely at sea. Shipbuilding technology has evolved rapidly every year, and many construction decisions “seemed like a good idea at the time.” Consequently, the ship constantly catches fire, with half of it under rebuild at any given moment.
The ship has sprouted catamaran-like secondary hulls, symbiotic growths connected by struts and pipes. For example, years ago, there was a plan to switch to trendy high-efficiency jet fuel turbines. During the switchover, we needed twice the fuel tanks, so we hastily bought a small jet fuel tanker, welded it to the hull, but then oil prices crashed, and the retrofit was abandoned. Now there’s a cute little jet fuel tanker stuck to the side, kept around “just in case.”
You’re darting over rope rigging when you encounter your current headache: the fuel pumping system. It was installed years ago by some assholes (it’s a well-known fact that once a sailor leaves a ship, he becomes an asshole happy to accept blame for any problem), and it uses custom parts from a long-defunct company, and they left no documentation. It works perfectly—unless someone tries to improve it. Every attempted tweak has caused a breakdown and a panicked fix. After too many disasters, other departments threatened to switch to electricity, so now nobody dares touch it.
This wouldn’t matter if fuel demand weren’t about to double with the new diesel generators coming online. The diesel generator team had been claiming that installation was imminent for over a year, but last week they started installing pipe fittings, triggering a scramble to figure out how we were going to double our fuel output.
Among the higher-ups, there’s been a big argument over whether to upgrade the existing system, build an entirely new system alongside the old one, or beg the diesel team to delay because we didn’t think they were serious.
Incrementally upgrading the current system is your best option (if it works): it’s proven reliable, cost-effective, and manageable. Solving this would be a huge feather in your cap. Never mind that three guys have already tried and failed. They didn’t have your skills!
You tell your boss, “I need a month to fix the bottlenecks.”
“Look,” your boss says, bags under his eyes from decades fighting battles that could only be fought but never won, “take three weeks. Just focus on writing a solid maintainability report. Anything else is biting off more than you—or anyone—can chew.”
After two weeks obsessively mapping every obscure pipe and valve placement, your brain is numb. You’ve discovered, to your horror, the fuel system is essentially a massive hydraulic computer, riddled with self-regulating siphons, inverted Pythagorean cups, and even what looked suspiciously like ancient water clocks. Pipes loop through unused chambers, occasionally rerouting fluid for no clear reason. Your best guess is it handles obscure pressure issues when the boiler starts up and while the outboard motors are on full power—but can’t test your hypothesis. And this is just one of the dozen baffling subsystems.
Your notes are a sprawling mess; you’re on version four of the “big explainer doc.” You dream about pipes, and you don’t like eating macaroni or penne anymore. You lean back heavily, rubbing your temples. There’s no way you’re solving this in the week left. You’ll come crawling back to your manager and go back to laying pipes for the pneumatic messaging system1, inching along toward adopting the standard everyone in the industry agrees is definitely the future.
Defeated, you go vent to your buddy from agriculture at lunch. You’re about to leave when he tilts his head.
“You know, I was doing some reading of some old docs, and a couple years back we had a project to increase fuel self-sufficiency. Middle East tensions spiked oil futures, so we’re probably still mixing in vegetable oils and ethanol from our wormy corn.”
You’re stunned. He continues, “Doubt anyone turned that off; boosts our budget.”
“Wait,” you say, dropping your spoon. “You’re making ethanol in the old fermentation tanks?”
“Yeah,” your buddy shrugs. “We’ve sent biofuel your way for maybe three years now. Nobody complained.”
You stare at the table. Suddenly, it makes sense: weird flow rates, unpredictable stalls, baffling chambers, strange edge cases.
You groan. You bet the work logs would confirm the new fuel system was built exactly as biodiesel inputs began. Those assholes who built the fuel pump weren’t assholes—they were heroes who unknowingly designed a system robust enough for viscous, unstable trash fuel. And just like you, they were probably engineers with no prior experience with fuel pumping systems. They just didn’t know any better. And you can’t even blame Agriculture—how were they supposed to know if nobody complained, and the captain’s goal of fuel self-sufficiency was achieved.
Your buddy squints. “So…should we stop?”
You exhale and put your head in your hands. A solution slowly comes to mind: disconnect the biofuels and hook up that unused jet fuel tanker. But that would take months and require routing pipes through medical. You winced. That was going to be a problem for your boss’s boss. After that, the fuel pump would handle the same throughput, but at much higher energy density. Then you could rip out the old pumping system entirely and replace it with something simple. A long-term fix.
“No,” you sigh. “Just tell me who to talk to so I can understand what’s going on at your end. Thanks for the tip.”
Later, you wonder if there’s a better way. Less chaos, better training, clearer planning, standards followed from the start.
But forty years ago, ships mostly hugged coasts, powered by sails and oars—essential yet small-scale. Now, enormous ships dominate the oceans, approaching “small sovereign country” status. The global economy became critically dependent on them in under two decades, shipping capacity exploding sevenfold. Everything is changing too quickly. Maybe in twenty years, shipbuilding will stabilize—or perhaps the chaos will only accelerate.