Conversation

This man literally has no idea what he's talking about. Stay in your lane, dude.

Quote Tweet

Jun 4, 2019

Thus, the testing burden is independent of typing. The number of tests you write and execute is unaffected by the type system of your language.

Show this thread

166

554

Replying to

His argument that type systems can’t reduce necessity of tests is ridiculous prima facie, but I do think there’s a valid point lurking in there about how type systems (i.e. proofs, i.e. mathematics) can’t fully displace empirical testing (i.e. science in the Popperian sense).

Replying to

and

This is why property-based testing is a thing

1:52 PM · Jun 6, 2019Twitter Web Client

Replying to

and

Sure, property-based testing can help catch many categories of defect that types alone do not: type system gaps, developer misunderstanding of type system, compiler bugs, etc. There’s also separate value in manually constructed (i.e. non-fuzzed) test cases; thus book reco above.

Replying to

and

One could argue that properties you can't encode in the type system are just deficiencies of the type system (and I'm not saying it's true, but just that it's not provably a test-only side of things)

Replying to

and

Sure, that’s the dream of dependent types: to make every desirable property of everything part of the type system. It remains to be seen how generalizable & practical that dream can be, but it’s promising! In the meantime, these various testing approaches fill the gap. But…

Show replies