Is Software Development Engineering?

HostThe voices in this episode are synthetic, and the script was written by a language model. The positions are real, and they come from the thread.

HostOver on a Mastodon instance called benno, there’s a debate about whether software development qualifies as engineering. It’s a short thread, but it gets to the heart of why we don’t have professional accountability in code.

HostIt isn't just about titles; it’s about whether legal liability would actually force the industry to stop shipping broken products. The room is split between those who want reform and those who think the industry is too heterogeneous to ever be standardized.

Liability will disciplineIf you want to be called engineers, you need the responsibility that goes with it. We’re sick of this ship it now, fix it later culture. Liability would change that overnight because suddenly, it’s not just a project manager’s deadline anymore. It’s a professional risk.

Not engineering tradeThat’s just borrowing prestige you haven't earned. Software development isn't engineering; it’s a trade. Engineering relies on model-based, mathematical construction. Nothing we do in software development works that way.

Not engineering tradeYou’re asking for liability in a field that doesn't even have a unified definition of what the work entails. You can't regulate what you can't define.

Liability will disciplineSo what? Maybe we need that structure to force the definition. If software engineers started saying I can’t ship this because it’s shit, that would change the power dynamic. It shouldn't be a trade if it’s building critical infrastructure.

Too diverse to standardizeThat doesn't work for everything, though. You can’t apply the same standards to a retail checkout system as you would to safety-critical infrastructure. The industry is just too vast for one code of ethics.

Liability will disciplineBut that diversity is the excuse we use to avoid accountability entirely.

Too diverse to standardizeIt’s not an excuse; it’s reality. High-stakes software already faces liability because it’s critical. If I wrote software that collapsed a bridge, I’d be liable. The market has tolerated faulty design elsewhere because it keeps costs down, and that’s a trade-off we’ve accepted.

Not engineering tradeExactly. You can’t impose engineering norms on a field that operates on economic tradeoffs rather than safety constraints. It’s not an engineering problem, and legal liability won't turn it into one.

HostThat’s where the thread seems to land—the belief that the market currently dictates these standards, for better or worse. One perspective that didn't get much traction was the idea that this should be treated like a vocational trade, taught in specialized schools rather than universities.

HostThere was also a grimmer view, suggesting that with LLMs, we’re heading toward a downward spiral of stupidity that no professional standard could fix.

HostIt’s a clean deadlock. The reformers want to use the law to force the professionalization they crave, while the others say the nature of the work makes that impossible. The question of whether we can ever standardize such a heterogeneous field remains entirely unanswered.

The same, as plain text