Skip to content

Comment on Dijkstra on the cruelty of really teaching computing science (1988)parent

Comments

A bridge that depends on all rivets being perfect is a bad thing. A proof that no longer holds outside of its domain of application is still useful.

The trick is to recognize when you need one and when you'd rather use the other.

Same in software - an application that fills a business purpose but has bugs or "errors" is correct.

An error free perfect piece of software hat doesn't is useless

AboutSource Built by g1lg1l

Hackerly is an independent reader for Hacker News, built on the public HN API. Not affiliated with Y Combinator.