July 6, 2026
Compile errors, but make it existential
"Software Engineering" Is Not Engineering
Coders just got roasted as commenters fight over whether software is math, engineering, or marketing
TLDR: An essay argued that making software is not true engineering in the same way building bridges is, because there’s no clear real-world test that ends every debate. Commenters split between defending coding as evidence-based engineering and mocking the whole field as math, marketing, or straight-up manipulation.
A spicy essay claiming “software engineering” isn’t really engineering tossed a lit match into the comments, and the crowd did not keep it civil. The article’s basic argument is simple: building software doesn’t work like building bridges or planes, because there isn’t one hard reality test that settles every argument. In plain English, lots of different coding styles can still get the job done, so people keep fighting forever about the “right” way. That led the author to suggest software may be closer to math than traditional engineering.
Cue the backlash. One commenter instantly sneered, “This is definably not engineering,” which pretty much set the mood: half grammar jab, half profession drag. Others pushed back hard, saying software absolutely is tied to reality because it runs on real machines in the real world. Another brought receipts, pointing to Dave Farley, who argues software engineering counts as engineering when it uses evidence, testing, and practical problem-solving.
But the funniest comments were also the meanest. One user joked that if software really worked like physical engineering, we’d have wallet-snatching robot claws everywhere and need Arnold Schwarzenegger instead of AdBlock just to go outside. Ouch. Then came the extra cynical twist: maybe software isn’t engineering at all anymore — maybe it’s economics plus psychology, especially when apps are built to manipulate users with sneaky design tricks. In other words, the article asked a serious question, but the comments turned it into a full-on identity crisis for the coding world.
Key Points
- •The article argues that software debates are hard to settle objectively because software design does not align cleanly with the testing model of science or physical engineering.
- •It describes traditional engineering as applied science whose models are constrained by physical laws and validated against reality or realistic simulations.
- •The article says software requirements can specify inputs, outputs, and runtime constraints, but much of the design space between them lacks clear objective metrics.
- •It claims many software approaches are Turing equivalent, so they can all implement the same clearly specified algorithm given sufficient resources.
- •The article proposes that software design is closer to mathematics because it involves constructing internally consistent abstract models rather than models tightly bound to physical reality.