April 10, 2026
Pretty code vs uptime: choose your fighter
Code is run more than read (2023)
Ugly code that works vs pretty code nobody runs — comments are wild
TLDR: The essay argues software should be judged by how well it runs in the real world and the cost of keeping it alive. Commenters split between calling for stronger rules to protect users, fatalism about big companies ignoring them, and war stories about ugly-but-reliable code that just won’t die.
A spicy essay claims the real cost of software isn’t writing it — it’s running it. Translation for normal humans: the hard part is keeping apps alive in the real world, and the bills for that keep coming. The author says ship it, keep it simple, and remember the business needs money, not just nice features. Cue the comment cage match.
One camp yells: protect users. A top comment argues that with strong rules, companies can’t squeeze customers just to please the budget, name-checking hated tech like digital locks and spyware TVs. Another camp goes full doomer: without a “crash or revolution,” big business will steamroll what users want anyway. Meanwhile, a finance veteran drops a confession straight from the vault: tons of “ugly” code runs flawlessly at elite firms — think duct-taped genius that never dies — proving reliability can live in a haunted house of hacks. The repair crew fires back with a car joke: yes, cars are driven more than fixed, but hiding the oil filter is a crime. And the cynics roast the latest trend: “crowdsourcing debugging” with machine learning? That’s a meme waiting to happen.
Verdict: it’s uptime vs elegance, user love vs revenue, and the comments are serving heat, gloom, and gallows humor.
Key Points
- •The article progresses from “code is read more than written” to “code is used more than read” and ultimately “code is run more than read.”
- •It emphasizes early and frequent user feedback to ensure software fulfills its intended purpose.
- •“Run” encompasses production operations such as deployment, upgrades, monitoring, auditing, fixing, and decommissioning.
- •It cites Dan McKinley to assert that long-term operational costs typically exceed build-time inconveniences.
- •Business constraints—budgets, marketing, deadlines, stakeholders, and revenue—must inform engineering decisions alongside user needs.