Is Haskell a good choice for software security?

Re:Is Haskell a good choice for software security? by the_af on Friday December 17, 2021 @11:52AM (#29594779)

I didn't claim there is a linear relationship. It's enough for me that there is a relationship. The OP asked "what's the use case" and I gave him several that ring true to me.

Whenever I write Python , I wish I was writing Haskell. No need to ask about artificial examples of linear algebra; just my boring day to day code in Python would benefit from having Haskell's type system. And yes, I do know about static typing for Python -- I find it's worse, tooling and effectiveness wise, than an actual static type system.

[ Parent | on HN ]

Re:Is Haskell a good choice for software security? by simiones on Friday December 17, 2021 @12:04PM (#29594939)

Well, GP was asking about Haskell vs Rust, not Python; and the discussion before was about Haskell vs Java as well.

And the example I gave is not entirely artificial - units of measure are often measured as an advantage of strong type systems, something that could potentially avoid the Ariadne disaster, and I find this common example unconvincing.

[ Parent | on HN ]

Re:Is Haskell a good choice for software security? by the_af on Friday December 17, 2021 @12:07PM (#29595000)

The GP also mentioned Python in another comment, and specifically asked "what about strong type systems", which makes the comparison to Python relevant.

> something that could potentially avoid the Ariadne disaster

Haskell or any language with modern static typing would definitely have helped. You wouldn't be mixing unit types and confusing them like it happened (if I remember correctly) with the Ariadne. Would it have been enough? I don't know! But it would have helped! Any measure of static and dynamic analysis would have helped for mission critical software, and it's evident the Ariadne had some serious deficits in this regard.

[ Parent | on HN ]