News for hackers. Stuff that matters.
Re:Seven deadly sins of talking about “types” (2014) by jnwatson on Saturday December 14, 2019 @09:49AM (#21789729)
And types can’t represent the properties shown by tests. Type checking as typically applied only catches the simple problems of feeding the wrong argument into a function.

Mere mortals can’t construct type systems to enforce most definitions of correctness or licenses or performance.

Re:Seven deadly sins of talking about “types” (2014) by naasking on Saturday December 14, 2019 @10:31AM (#21789944)
> And types can’t represent the properties shown by tests.

That's not true generally. Dependent types can be used to represent any property that a test might check, and lightweight alternatives can prove a subset [1]. And this makes sense from the type of logical property being established: you can easily turn a universally quantified proposition into an existentially quantification, but going the other way around means checking all possible permutations of values.

However, proving a universally quantification is sometimes (often?) far more difficult than proving an existential quantification, which is why programs have to be structured in a certain way to make them amenable to such proofs.

> Mere mortals can’t construct type systems to enforce most definitions of correctness or licenses or performance.

Type systems always improve performance, so I don't know what you mean by that exactly.

[1] http://lambda-the-ultimate.org/node/1635

Re:Seven deadly sins of talking about “types” (2014) by cobbzilla on Saturday December 14, 2019 @10:03AM (#21789784)
It’s not an either/or. Both types and tests help, for different reasons.

For programming in the large, both should be used. In the long run it’s about saving time that is otherwise wasted on wolf-fencing and hair-pulling; and in getting new developers up to speed faster.

Re:Seven deadly sins of talking about “types” (2014) by jcora on Saturday December 14, 2019 @10:11AM (#21789826)
Actually with dependent types you totally can encode proofs that your logic is correct. But I don't really see that trickling down into mainstream programming anytime soon.